Papers of the British School at Rome http://journals.cambridge.org/ROM

Additional services for Papers of the British School at Rome:

Email alerts: Click here Subscriptions: Click here Commercial reprints: Click here Terms of use : Click here

The Classical Topography of the Roman Campagna.—III. (The Via Latina)—Section II.

T. Ashby

Papers of the British School at Rome / Volume 5 / January 1910, pp 214 - 425 DOI: 10.1017/S0068246200005316, Published online: 09 August 2013

Link to this article: http://journals.cambridge.org/abstract_S0068246200005316

How to cite this article: T. Ashby (1910). The Classical Topography of the Roman Campagna.—III. (The Via Latina)—Section II.. Papers of the British School at Rome, 5, pp 214-425 doi:10.1017/S0068246200005316

Request Permissions : Click here

Downloaded from http://journals.cambridge.org/ROM, IP address: 134.153.170.22 on 15 Mar 2015 PAPERS OF THE BRITISH SCHOOL AT ROME.

VOL. V. No. 6

THE CLASSICAL TOPOGRAPHY

OF THE ROMAN CAMPAGNA

PART III

SECTION II

BY T. ASHBY, D.LITT., F.S.A. Director of the British School at Rome; formerly Craven Fellow in the University of Oxford ; Corresponding Member of the German Iniperial Archaeological Institute..

LONDON: 1910 THE VIA LATINA (SECTION II.).

INTRODUCTION.

THE work of preparing this, the second portion of the description of the classical topography of the Via Latina, has shown me more clearly than before how impossible it is to hope to attain finality in dealing with the Campagna. The late Henry Stevenson's notes now in the Vatican Library (those which especially concern this district are to be found for the most part in Vat. Lat. 10572) are a perfect mine of information, especially when taken with his own copies of the Staff Map on which the ruins he found are marked (now bound up together as Vat. Lat. 10587 B), and one realizes more than ever the value and extent of the work he might have done had he lived longer. As I have examined them carefully, I. have given full details of their contents. The maps for the present volume were unfortunately made before I had time to consult these valuable sources of information. And yet, when I came to go over the ground again, I found that there were many ruins that even he had not noticed, some of them of considerable size and importance. The truth of course is, that in hilly country so shut in by enclosures, covered too in the main by vineyards, oliveyards, or gardens, and wooded in the higher parts, it is impossible to get the clear distant views that are obtainable in the open Campagna; and even then one cannot be sure, without actually passing over every bit of ground, that there are not some ruins beneath the soil the presence of which is only disclosed by debris. One comes to realize more and more how thickly populated was this part of the country, which seems to have been the favourite summer resort of the wealthier Romans. Much more research in archives, too, would have to be undertaken before the material available was anything like exhausted. In the present section I have used to some extent the reports of excavations in the Atti del Camerlengato, Titolo, iv (1824-1854) now preserved in the 216 THE BRITISH SCHOOL AT ROME.

Archivio di Stato : the reports for the succeeding period in the records of the Pontifical Ministry of Agriculture, Commerce, and Fine Arts (1854-1870) preserved in the same archives, I have not as yet touched. The former contain, as will be seen from some of the specimens I have given, most valuable information, of which I hope to make further use.1 On the other hand, besides Prof. Tomassetti's work on the Via Latina already cited, I have had the advantage of being able to use an excellent little book by Father F. Grossi-Gondi, S. J., // Tuscolano mil' eta dassica (Rome, 1908), with illustrations from photographs, a good map, on the • same scale as the Staff Map and my own (1:25,000) and a full bibliography,2 which has appeared since the publication of vol. iv of the Papers. A residence of several years at Mondragone has given him the opportunity of accurate local study, and the results have been given to the the world both in articles in the Bullettino Comunalej&nd elsewhere, and in a work on Mondragone itself (La villa dei Quintilii e la villa di Mondragone, Rome, 1901). The present book is an excellent handbook.

1 The addenda to this and the former parts of the Classical Topography of the Roma Campagna are postponed owing to considerations of space. 2 In regard to the bibliography I may notice the following points, (i) The MS. cited as. Anonimo, Viaggio Antiquario in aletme citta del Lazio. Osservazioni su Tuscolo, and as having been sold in the Vespignani sale as No. 106, is in reality an inaccurate reference to the notes of Nibby (No. 581 in that sale) now in my possession (cited as Schede in the text). (2) The album Veteris Latii Anliqua Vestigia, Rome, 1751 is entered twice—once under Anonimo, once under A16 Giovannoli. I do not know of there being any ground for the latter attribution. Almost all the plates as a fact are identical with those in Corradini and Volpi's Velus Latium Profanmn (1704-45). The Veteris Latii antiquitatum amplissima collectio, noted under the year 1771 (really 1776), is a second enlarged edition of the same collection. (3) 'Domenico (Fra) MS. della Biblioteca del Seminario Vescovile di Frascati.' This is the same MS. as that quoted by Lanciani in Bull. Com. 1884, 172 sqq. (Cod. tusc. 14, I. 11, Antichita del Tuscolo e descrizzione del Lazio eseguita da P. Domenico Cappuccino da Frascati). My copy of the Bull. Com., which belonged to Stevenson, contains additions made by him to Lanciani's copies, so that I have not thought it necessary to re-examine the MS. myself. It must have been written towards the end of the seventeenth or the beginning of the eighteenth century (infra, 243), as the passage as to the Villa at Kontana Piscaro (Lanciani, loc. cit. p. 201) is an exact copy of Kircher, Latium, 73 (published in 1671). (4) The exact title of the views by Labruzzi which were engraved by Parboni and Poggioli is Vedute ed avanzi delf antica citta di Albalonga ora Albano disegnali dal vero etc., and there are 24 drawings, not 26. None of these views actually relates to Tusculum ; and I do not find any entry in the Catalogue of the Stourhead Library (London, 1840) pp. 543 sqq. which would justify the supposition that Labruzzi or Sir R. Colt Hoare did any drawings there. (5) To Piacentini's works we should add Commentarium Graecae Pronuntiationis (Rome, 1751) and Be Tusculano Cicercnis nunc Crypta Ferrata (Rome, 1758) cited by Venuti in the preface to, the Monumenta Mattheiana, p. iv, No. 2. (6) Cozza's work // Tusculano di M. Tullio Cicerone first appeared in Giornale Arcadico, cxc. 97 m- CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 217

XIII.—THE VIA LATINA FROM CASALE CIAMPINO TO GROTTAFERRATA {from the Tenth to the Twelfth Mile). In the cutting by which the tramway leaves the highroad the line of the Via Latina was discovered 40 metres within the vineyard, running 300 E. of S. On each side of it remains of tombs of opus quadratum and opus reticulatum and later burials under tiles were found: a terra-cotta sarcophagus was also discovered. The interval available for the road, including footpaths, appeared to be about 8'5o metres, but the actual road- way, here as elsewhere, was probably about 4/20 metres (14 Roman feet). Here I found some brickstarhps on tiles of the first century, perhaps used, however, for late tombs a capanna—C.I.L. xv. 1383 (a or c), 2333 a (?) (IVN ) and a fragment of 169 b (?) (OP . D .. ) the letters being larger than usual, with a dog (?) in the centre. In the fieldwall, before the tramway was made, I copied a fragment of an inscription on a marble slab O'2O metre thick in well cut letters

There are, as we saw in Papers iv. 130, in the Vigna Gentilini and the Vigna Costanza Senni, the next vineyard to it to the S.E., the remains of four tombs above ground, two on each side of the course of the ancient Via Latina. The first, on the right of the road,1 though I have wrongly marked it on the left, is a circular mass of concrete: then comes another one, on the left, of concrete, preserved to some height, which is square, and then two others, less well preserved, on the left-hand side of the road.2 Many ancient gems are said have been found here (Stevenson, Vat. Lat. 10572, 34V). 1 The second part of Lanciani's article, describing this section of the road, has not yet appeared. The map is reproduced in Wanderings in the Roman Campagna (London, 1909), p. 23. 2 The entrance to the catacomb mentioned in Papers, iv. 130 is situated immediately to the S, of the tram line, which indeed cuts through some of its galleries. It has been re-opened, and will shortly be carefully explored. At the entrance we saw a marble slab with the following inscription :— oMirnuniAE vm\tRiAt, A . V . XIII . M . V PERMISSV SEMPRONIAE DIGNITATIS OPTIME FEMINAE (sic) !' H. T. D . M. AB ESTO .218 THE BRITISH SCHOOL AT ROME.

On the S.W. side of the road, where a reservoir is marked in the map, are the remains of a villa with two platforms : on the lower is a reservoir, which originally 'had four chambers (three of which are preserved) intercommunicating by means of five arches in each of the intermediate walls. The back wall of the substructions to the N.W. is faced with roughly horizontal masonry in selce, the blocks having smooth faces and joints. The situation of the vineyard of Giovanni Conti, son of Nicola Conti, of which Stevenson (Vat. Lett. 10572, 57v) speaks as being the third going down (from Grottaferrata) before reaching the Vicolo di Mola Cavona, is clear from ibid. 101; it must be that in which is this cistern, on the S.W. of the Via Latina. There were discovered, Conti informed him, many lead pipes and a statuary group of marble of a male and a female figure embracing one another, not a Cupid and Psyche, for they were not winged ; nor was it the same as the Cupid and Psyche or Venus and Cupid, of porphyry (f. 101) found in the Vigna Enrico Conte fu Nicola, which is the last, and is bounded by the Vicolo di Mola Cavona but not by the Via Latina. A branch road from the Via Cavona, apparently going towards the reservoir, was also found. To the S.E. of the group of tombs above described, on the N.E. edge of the ancient road, are the remains of a very large villa, which, as we have seen {Papers iv. 128) Grossi-Gondi assigns to the Vinicii Opimiani (cf. also his Territorio Tuscolano, p. 53, with the view of the lower terrace on tav. iv.). It is in two terraces; the supporting wall of the lower one, cut through by the electric tramway line, had originally eleven arched recesses on the N.W. side (some of which are shown in Grossi-Gondi's photograph) each with a span of 4/90 metres and a depth of S'lO metres. Above them is a horizontal line in the concrete, and, after another layer, a fiat surface (not apparently a floor) is reached, bounded towards the front of the terrace by a low wall, and towards the back by walling in two thicknesses of oyo and C47 m. respectively. The intervening space, a

(Semproniae Veneriae a(nnos) v{ixit) xiii m(eases) v permissu Semproniae Dignitatis optimae feminae: h{oc) t(umulo) d{olus) mialus) abesto.) The slab is broken at the top : it is C40 metre wide, and the remaining portion of the tablet on which the inscription is cut is O"i6 metre high. Below it the slab goes on for o'3i metre: and on each side is a small rectangular part cut out, as if to fix it better in its place. The letters are 0-02 metre high : on the plaster of one of the loculi on the left going down are scratched the letters CVRII. Outside the catacomb (but no doubt brought from it) we saw the following brick- stamps on flange tiles, C.I.L. XV. 223 a (about 14OA.D.), 369 (148 A.D.). CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 219

little less than five metres wide, is at present filled with earth, as the photograph shows ; and the object of the arrangement is not very easily comprehensible. On the north-east side of the platform are no recesses, but a wall supported by buttresses in opus retiailatum, while on the south- west, towards the road, no supporting walls were necessary, and on the south-east is the wall of the upper terrace. On the lower terrace the work of laying the tramway brought to light one intermediate wall, parallel to the front wall. Several brickstamps were also found. I noted there a fragment of the stamp C.I.L. xv. 563 (PAETIN . . . . M VINIC PA . . .) of 123 A.D., on a brick which may have belonged to one of the pilae of a hypocaust, a fragment of ib. 1196 (123 A.D.) and a fragment oiib. 1332 (?) (L N/EV—first century A.D.) On this platform there were also, it seems, burials of a much later period, the bodies being placed under tiles ; on some of these were the stamps C.I.L. xv. 454 c (123 A.D. several copies) 1.121 c (first century A.D.), 1318 b (P N^EV first century A.D.). The upper terrace wall of the villa is nearly 70 metres back from the lower; it is built of opus incertum of selce of somewhat large pieces, measuring about C25 by 0*22 metre, and behind it runs a vaulted passage 1 metre wide. Both these two lower platforms are orientated with the Via Latina, being at right angles to it. On the upper platform is a ruined vineyard house, in which the ancient pavement of opus spicatum has been used. There are also several blocks of peperino, and two Corinthian capitals, one belonging to a rectangular pilaster, the other to a round column, in the same material. The flange tiles of the roof had spouts in the form of lions' heads. Higher up is a third platform, perhaps belonging to the same villa, and faced in the same way as the second, but on a different orientation (though this is not indicated on my map, nor on Lanciani's) and not coming so far to the N.E., so that the tramway line does not actually cut through it. On the same orientation and level with it is a very large water reservoir, with two vaulted chambers, with the usual round air-holes in the roof; each chamber is 5870 metres in length and 2'62 metres in width. The wall which divides them, O"92 metre in thickness, has arched apertures in it, alternately 1-15 metre and 0-59 metre in width and •O'9O metre apart. The former go down to the floor level, which must be some two metres below the present ground level, making the total 220 THE BRITISH SCHOOL AT ROME. height to the spring of the vault above about 5 "30 metres ; while the latter are only r8o metre in height, their rounded tops coming within one metre of the spring of the vault. To this, I think, De Simoni {Lettere famigliari, Rome, 1831, p. 6) is alluding, though his description does not make the identification certain. Below this reservoir are the remains of a peristyle in opus quadratum of peperino, which was cut through by the tramway : a line of gutter slabs o-47 metre wide, with flat slabs going parallel to them, was found orientated about 30° E. of N. A little to the N.W. in the tramway cutting are remains of underground passages for storing rain-water, lined and floored with cement, which were about O'o,o metre in width,, and originally some 1*50 metre in height. The type is a common one in the Campagna. A round shaft C93 metre in diameter found on the S.W. side of the cutting may have communicated with them. To the N. of the peristyle was the pavement of a road going about due E., and descending sharply into the valley, where it probably joined a road, marked also in Lanciani's map, which ran parallel to the Via Latina, below these villas, from the Via Cavona to the Villa Montioni. In the other direction it turned sharply to the S.W. and, crossing the villa, reached the Via Latina. An allusion to the discovery of this road is probably made in Not, Scavi, 1905, 244/ where it is described as about 5 metres wide and as crossing the tramway line obliquely from E. to W. (sic). On the S. of it were found walls in opus reticulatum, only one chamber being measurable : it was 3-50 metres long and 2-io wide. Close to these walls was a circular shaft, walled, 1 metre in diameter, communicating at the bottom with an underground passage (no doubt the one we have just mentioned). Otherwise the official reports on the discoveries made in this district are lamentably scanty. The somewhat detailed description I have given is based on frequent visits during the progress of the works. The tramway line (not indicated on my map) passes between the twa groups of ruins to the S.E. of the large villa just described. These two groups probably belong to another large villa. That to the N.E. is a plat- form of opus reticulatum in selce and tufa, orientated 25° W. of N. Within

1 The fact that it is described as being some 680 metres from Villa Senni shows that Grossi- Gondi (p. 60) is wrong in attributing this description to one of the roads above the Villa Montioni. It is, however, a great pity that any ambiguity should have been possible. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 221 it, at its lowest level, is a subterranean reservoir measuring 2^97 by 416 metres, with vaulted roof, and running E. and W: it has a circular shaft in its N.W. angle. To the S.W. are large vaults with brick facing, orientated about 10° E. of N. In the tramway cutting were the possible remains of a late burial ' alia cappuccina.' To the N.E. again, and on the further side of a depression, is the prominent platform of a very large villa, just to the S.E. of which is a house marked 214. The platform is faced with opus incertum of selce, approximating to opus reticulatuni, in bands about 2 feet high. On the N.E. there is a wall with a lofty projecting tower at the N. angle (with a vaulted chamber inside it), on the N.W. a high substruction wall at first, and then as the ground rises, low arcades, some twenty in number. On the S.W. there is only a little walling, the edge of the platform being mainly formed by the natural rock. Below it, to the N.W., another villa is marked in Lanciani's map, but there are no traces of it to be seen : it is quite possible, but by no means certain, that a modern cultivation terrace marks the site of an ancient terrace wall. At the electric power station for the tramway I saw the following objects, found in this section of the tramway works : part of a draped statue, an ox-skull in marble about 0*50 metre high (now in the museum of the Abbey of Grottaferrata), various fragments, several terra-cottas, and a few brickstamps, C.I.L. xv. 479 (123 A.D.), 1121a (ist century), 1239 a (ist century), with points between some of the words—thus Q • LEPIDI • O • F IDVARI : and a rectangular stamp, unpublished in C.I.L., MARIO, in one line only—not therefore a fragment of 818 L. Antonius Mario. Opposite the uppermost of the villas we have described, on the right- hand side of the high-road, just before the fifteenth kilometre stone, a lane turns off to the right, which is probably ancient: no pavement is to be seen in situ, but there are paving stones and fragments of selce in the field walls. It is supposed by Rocchi to be the deverticulum leading to the springs of the Aqua Tepula: in vol. iv. p. 131, 1. 18. I have wrongly attributed to him what is really Lanciani's view. The modern house called La Torretta commands a fine view, and probably occupies an ancient site: no platform of a villa is visible, but there are a few blocks of marble and other debris. The lane descends past it steeply, and the ancient road is believed by Lanciani to follow the same 222 THE BRITISH SCHOOL AT ROME. course, while Rocchi and Grossi-Gondi, whom I have followed in my map, take it along the side of the valley, and then make it descend only when it reaches another path, which is almost opposite to the springs of the Aqua Tepula, being at the same place joined by the path coming from the Camposanto of Grottaferrata {infra, 227). No pavement is to be seen in situ in either case, but Rocchi (Diss. Pont. Ace. Arch. Ser. II. vol. vii. (1900) 224 n. 3) notes the discovery in 1898 of some pavement belonging to it in a vineyard about 100 metres N.W. of this junction of paths. The source of the Aqua Tepula has been identified, and no doubt rightly, with the Sorgente Preziosa ; but the credit of the first identification is not due to Secchi, as Lanciani {Comentari di Frontino cit. p. 294) and Tomassetti (Via Latino., p. 84 n.) suppose, but to a far older authority, as is clear from the passage of Holste quoted in the footnote.1 The temperature of the spring has been observed to be 6i°-63° Fahren- heit in winter, when that of the air was only 47°, while that of the Aqua Julia {infra, 386) was 5o°-52°. This circumstance, and the agreement with the indications of distance given by Frontinus, suffice to identify it; but no remains of its channel are to be seen, or have ever, so far as I know, been found, until it appears near Le Capannelle, running above the Marcia and below the Julia. There are remains of aqueducts near the Mola Cavona and in the Valle Marciana, which I agree with Tomassetti (p. 87) in considering to be mediaeval. The valley itself was according to M. S. De Rossi the bed of an ancient lake, on the banks of which Latin pottery has been found (cf. Primo Rapporto sidle scoperte paleoetnologiche 41; Secondo Rapporto (1867) 30). Of its classical name we have already spoken {Papers, iv. 126): in the tenth century we find a church of S. Maria in Diaconia mentioned there in

1 Holste, Cod. Dresd. F. 193, f. 43. 16 October, 1649. Inspexi fonlem Tepulae, vulgo nunc la pretiosa dicta ; est in valle Marciana sub Burgetto castello diruto in via Latina ad XII lapidem ;. in dicta valle ad Crabram est officina ferraria, ultra earn ad CCC circiter passus scaturit fons aquae copiosissimus, vulgo La Pretiosa dictus, quam Tepulam esse ex Frontino certissimum est, distat enim duobus m. pass, a decimo (vulgo Le Murene) dextrorsum deflectentibus. Sed cum Frontinus neget Tepulam certum habere fontem sed ex venis collectam, existimo vena's illas in unam corrivatas, postquam Juliae ductu receptae (43V) amplius in urbem fluere desierunt, hoc fonte simul prorupisse. Quod etiam idem Frontinus Tepulam agro Lucullano concipi ait, id huic fonti maxime convenit. Nam villae Luculli maxima extant vestigia sub Burgetto ad sinistram viae Latinae, ubi substructions ingentes per vineas aliquot porrectas inspexi (the reference is to the ruins described supra, 218) : ab hisce vestigiis villae Lucullanae DCC circiter passibus abest fons ille Fretiosa dictus; puto tamen multo longius se protendisse agrum Lucullanum per subjectam planitiem usq(ue) ad pontem Crabrae sub Decimo ubi immensa illius villae vestigia visuntur quae vulgo il Centrone dicuntur. CLASSICAL TOPOGRAPHY or THE ROMAN CAMPAGNA.—III. 223 bulls of 95 5 and 962, another form of which, Jaconia, still occurs as the name of the valley according to G. B. De Rossi {Bull. Crist. 1872, 101 = 113 of the French version). What may be the origin of the name Valle Nicosia, which appears on the map as the name of the small valley N.W. of the Colle dell' Asino, I do not know. The discovery of a fragment of an unimportant funerary inscription in the district of Valle Marciana, but in the territory of Marino {i.e. probably a little to the N.E. of Le Selve, and certainly W. of the communal boundary which runs southwards from the Colle dell' Asino) is recorded in Not. Scavi, 1903, 22. It had, however, already been copied by Stevenson {Vat. Lat. 10572, 2). In 1840 the remains of a small temple measuring H'^,8 metres long by 6-36 wide, with Doric columns, were discovered on the right bank of the Marrana Mariana in the Valle Marciana, upon a small hill of peperino, and are described by Blessig {Bull. 1st., 1840, 161 = Canina, Tuscolo, 99 : the passage is quoted in full by Lanciani, Bull. Com. 1905, 143). The walls were of concrete, and at intervals there were bonding courses of tiles, one of which bore the stamp C.T.L. xv. 725 (Faustina the younger). It contained a dedication to Septimius Severus by the people of Tusculum, and ran thus [Divo] Severopatri Antonini Pit felicis Aug{usti) \Tii\sculan\i\. Cf. C.I.L. xiv. 2497, where it is wrongly inferred from Blessig's account that the discovery took place a mile below the abbey of Grottaferrata: what is really stated is that the valley of the Marrana widens out at that point, and forms what is known as the Valle Marciana, but the precise site of the discovery is not given, and as Lanciaui says, cannot exactly be fixed. On September 6th, 1597 a license was given to Marzio Colonna to excavate in a place called Valle Marrani in the territory of Tusculum, and to extract any marble, travertine, statues, or treasure that he found. (Provvedimenti del Camerlengo 1597/98 c. 149, in the Archivio di Stato at Rome, quoted by Lanciani, Storia degli Scavi, iii. 50.) The ruins actually visible in the valley are entirely mediaeval (Tomassetti, p. 87) and the same is the case at the picturesque waterfall called the Cascata dei Gavotti. Indeed, I have not been able to find any traces of the specus of the Aqua Julia either in the Valle Marciana or in the gorge going up to the Ponte Squarciarelli (cf. Not. Scavi, 1887, 82). The mediaeval castle of Borghetto, which is built right across the Via Latina, and seems to have been intended to block it, was probably erected 224 THE-BRITISH SCHOOL AT ROME. about the tenth century by the Counts of Tusculum. Like that of the Caetani on the Via Appia, it led, as Nibby justly remarks (Analisi; i. 300), to the abandonment of the road, which in both cases had to pass right through the castle. It is mentioned under the name of Civitella in the bull of 955 of Agapitus II. in favour of the monastery of S. Silvestro. In 1436 it belonged to the Savelli. Tomassetti (p. 133) says that it rests upon foundations of selce concrete of the Roman period, belonging to a villa ; but these foundations are in reality mediaeval, and follow the whole line of the mediaeval enceinte. Nor is the cistern nor any other of the constructions within the walls anterior, in my opinion, to the Middle Ages. De Rossi, Bull. Crist. 1872, 117, places near here (there are several villas, of course, which would suit the identification) the villa of the Javoleni, C.I.L. xiv. 2499, a dedication to C. Iavolenus Calvinus Geminius Capito having been found in 1741 not far off (' nella via Latina presso il Castellaccio' according to Giorgi). His career included a tenure of the consulate as suffectus (the date is uncertain), but his name is not elsewhere mentioned. (Prosopographia, ii. p. 151, No. 12). In the donation of the seventh century made by Sergius I. to the church of S. Susanna (compare the register of Gregory II.) the fundus Capitonis cum casis et vineis seu oratorio sanctae Faustinae posito via Latina rnilliario plus minus XII iuxta massam Marulis is mentioned, which must be identical with the site of the villa of the Iavoleni and seems to place it rather further along the road. In the Massa Marulis there was also a Basilica of S. Peter1 and a colonia quae dicitur Pofinis situated just behind its apse. The account of Ramagini states that the inscription was found'nel territorio di Grottaferrata (that is on the N.N.E. side of the Via Latina) rimpetto alia vigna di Monsignor Ciampini' which would agree sufficiently with Giorgi's account if we suppose the inscription to have been found just below Borghetto, ' in sight of' or ' opposite to ' the vigna Ciampini. I do not think therefore that Grossi-Gondi {Bull. cit. 28) is right in emphasizing any slight disagreement between the two accounts ; and the two sepulchral inscriptions which he cites do not prove very much one way or the other.

1 From two documents of 955 and 962 we know that this church was deserted, but that its ruined walls still existed near the Valle Marciana (De Rossi, Bull. Crist., 1870, 106 ; 1872, 117). In the notitia fimdorum of the church of SS. Giovanni and Paolo on the Caelian we find two estates mentioned near the eleventh milestone—Fundus Publica and Fundus Casa Quinti—which we cannot fix more exactly. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 225

One of them {C.I.L. xiv. 2546) belongs to a freedman of the lavoleni, one L. Jabolenus Onesimus, and was seen in the Vigna Ciampini by Lesley; the other {C.I.L. xiv. 2546 a) was erected to one Iavolena Artemisia and was found in 1885 in a vineyard between the Oliveto Porcacchia and the road to Frascati {i.e. the road to the so-called Torrione di Micara) but not, as is wrongly stated in Notizie degli Scavi, 1885, 77 (the statement is not, as Grossi-Gondi says it is, repeated by the Corpus, which merely places the discovery ' in the vineyard called Borghetto') in the territory of Marino, but in that of Grottaferrata. In the same vineyard was found the unim- portant sepulchral inscription C.I.L. xiv. 2541a. Lanciani {Bull. Com. 1884, 189) identifies the villa of the lavoleni with the large villa under the Vigna Montioni. This identification, as Grossi-Gondi notes, agrees with the indication given by the donation of the 7th century, but not with the provenance of C.I.L. xiv. 2499. At Borghetto Mr. Baddeley has found the brickstamps C.I.L. xv. 690 (Severus—the figures in the centre of the stamp being cancelled, as in other cases), 1330 (1st century A.D.), and Giorgi notes as found at Borghetto in 1732 the stamps ibid. 1104 (end of 1st century A.D.) 1800 (Septimius Severus(?)). There was also found, in 1885, the lower part of a seated Egyptian statue, in black basalt, used as building material in a foundation wall of the castle {Not. Scavi, 1885, 159). The upper part, with the head, was found in 1900, but secretly sold to a foreigner (Tomassetti, Campagna Romana, (Rome, 1910) i. 82 n.). Close by was found a cippus with an unimportant sepulchral inscription {C.I.L. xiv. 4230) and a fragment {ibid. 4230 a). In a vineyard near Borghetto another unimportant sepulchral inscription was found in 1894, and now forms part of the epigraphic collection of the Abbey of Grottaferrata {Not. Scavi, 1894, 380). Just beyond the castle falls the site of the nth milestone. De Simoni, Lettere Famigliari, (Rome, 1831) p. 7, notes above the castle of Borghetto on the left a square tomb of selce concrete measuring 14^50 on each side outside. The internal chamber is barrel-vaulted, as are its two entrances, and measures 5'35 m. by 4*20 m. In the centre of the vault is a hole 0-36 m. in diameter, penetrating through the vault, which is O'6o m. thick according to De Simoni, but really about 1 m. In his measurements he omits the vaulted passage or approach on the S. side, 4'2O m. long by 2-io wide. This building is indicated in our map. Q 226 THE BRITISH SCHOOL AT ROME.

To the N.E. of it is the point at which the (probably) ancient road mentioned on p. 220 would terminate, reaching a large villa,1 with a well marked and lofty platform, the N.E. side of which is occupied by the buildings of the Villa Montioni. The modern path indeed cuts across the platform of the villa, but the ancient road must have stopped at the lower platform. Of the substruction walls but little is1 preserved: a lower platform wall of opus reticulatum, like the rest and on the same orientation, still remains, and here may be noticed fragments of paving stones, possibly from this branch road. The tramway a little higher up passes across some vaulted chambers belonging to it, in opus reticulatum with stone quoins, some of which were water-cisterns. To the S. of the villa is another tomb, and to the S.E. an ancient road goes N.E. to join the road which ran up past the Torre di Micara to Frascati {infra, 244). To the S.E. of it is yet another tomb (?) marked by Lanciani, but omitted in my map, the concrete foundations of which are preserved. To the E. of it is a huge reservoir, with eleven chambers, each 16 metres long and 370 metres wide : the partition walls are cr6o metre thick, the outer walls O'o.o metre thick, and there are also external buttresses. This reservoir is just to the N. of the 16th kilometre of the modern road : on the opposite side of the road, a little further back, on the hill-side overlooking the Valle Marciana, are scanty remains of a villa. To the E.S.E. of the large reservoir another ancient road, which had not been followed far, was said to exist in the vineyard. Freshly found paving stones were certainly visible, and I marked its direction as best I could from the information given me. It was not found in the cutting of the tramway, the work for which, however, revealed the existence of a drain, cut in the rock, about two metres high and half a metre wide. To the E.S.E. of the road last mentioned is another smaller reservoir (6 chambers) ; and here Lanciani's map ends. We now reach the path which intersects the Via Latina just W. of

1 Nibby (Analisi, iii. 357) notes the existence in the place called Porcacchia between the Torre di Micara and Borghetto of the fine substructions of a villa, with two terraces one above the other, the lower having rectangular niches, the upper alternately rectangular niches and plain walls : the whole was constructed of chips of selce. The reference may, I think, be to this villa. C.I.L. xiv. 2564 (a marble fragment with the letters . . iolani) was found in the Vigna Rosati in Cardoni's time (1757) not far from Ciampino and about a mile from Grottaferrata. He mentions, there the remains of a large villa, with walls of opus reticulatum and the remains of a road : here he noted the brickstamps C.I.L. xv. 595a, 10 (Hadrian) 2244, 2267 (both first century A.D.). It is very possible that this is the villa of which we are speaking. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 227 the Camposanto (cemetery) di Grottaferrata. It is of ancient origin : its pavement, 250 metres wide, was discovered during the construction of the electric tramway (though its direction is inaccurately given as from east to west in the reports) {Not. Scavi, 1905, 244) and some of it may be seen in situ. Following it to the N.E. we find another path diverging from it in a N.W. direction, which is, to judge from the existence of paving stones, probably also of ancient origin, at least as far as the cross road going N.E. from Villa Montioni ; but beyond that it presents no definite traces of antiquity, being paved with pieces of selce, which show no signs of ever having been parts of paving stones. To the E. of the house at point 236 are three small dots in the map : these indicate a large reservoir with five chambers intercommunicating; by means of arches. It may have supplied the villa above Fontana Piscaro. The main path goes on across the Macchia di Grottaferrata in a N.E. direction, and to the W. of the Villa Muti joins the road from the twelfth mile of the Via Latina {infra, 239). Returning to the Camposanto and following the path to the S.W. we find no certain traces of its antiquity, though it is a prolongation of what is undoubtedly an ancient line, and falls into the path mentioned on p. 222. Two groups of debris, marking, perhaps, the sites of ancient villas, are indicated to the E. of it in the map ; while to the W. are remains of another villa. The locality bears the name of Bagnara {infra, 256). On the N. side of the high-road we must place the Vigna Conti to which Stevenson refers in his notes (Vat. Lat. 10572, 57 v) as belonging to Giovanni Conti and being near the lane opposite the cemetery: in it the ancient road and two cinerary urns of peperino had, he says, been found. A little further on, to the E. of the Casa Santangeli, is the Vigna Giammarioli, where {ibid. 34) mosaics, water-pipes, and sculptured marbles had been discovered. Close to the Casale Santangeli we must place the point at which diverged an ancient road, which Lanciani {Comentari di Frontino, cit. 296),. Rocchi {Diss. Pont. Ace. Arch. Ser. II. Vol. vii. (1900) 223 sqg.) and Grossi-Gondi (p. 60) identify with the deverticulum mentioned by Frontinus (i. 9) as diverging near the twelfth mile of the Via Latina, and leading to the springs of the Aqua Julia, which it reached after two miles more. The line as given on my map is that of Rocchi but may be Q 2 228 THE BRITISH SCHOOL AT ROME. incorrect.1 Grossi-Gondi points out that if, as Lanciani does, one supposes the deverticulum to run in a straight line, it is barely ij mile (and not 2 miles) long as far as the springs (which are at the Ponte Squarciarelli, infra, 386). This, it is true, might matter little : for Frontinus does not aim at strict exactitude in these indications ; but, further, traces of an ancient road were observed in the first half of the nineteenth century near the Fontanaccio2 (marked on our map to the. N.W. of the Abbey of Grottaferrata) running in a S.E. direction, and other traces of it have been noticed near and even within the Abbey itself more recently. Beyond the Abbey it reappeared in the Vigna Santovetti (Secchi, Intorno ad alcune opere idrauliche antiche rinvenute nella Campagna di Roma, p. 35, from Atti del Nuovi Lincei xxix. (1876)) and must have rejoined the line followed by the modern road near the Mola, and so have reached the Ponte Squarciarelli. Grossi-Gondi shows its probable course clearly on his map, and also discusses the question whether this road was private or not, coming to the conclusion, as against Rocchi, that it was not, partly owing to the discovery, in the garden of the Abbey, of two unimportant 3 sepulchral inscriptions {C.I.L. xivc 2524, 2559) and possibly of an actual tomb, and of cinerary urns in front of the entrance to the church in 1903 (pp. 61, 80 n. 5). Later tombs covered with tiles (three of which bore the stamps C.I.L. xv. 534.2, 581.n,4 1081.9) and containing lamps were found in the Vigna Passerini, about forty paces from the Abbey, in 1735 (De Rossi, Ann. Inst. 1873, 207). The Vigna delle Monache (formerly Villa Carbone and Villa Beccari) lies to the N.W. of the Abbey : in it are the remains of an ancient villa, in two distinct parts. One, to the N.W., lies under and to the S.W. of the house marked 321 in the map, which was the Casino Carbone in Stevenson's time {Vat. Lat. 10572, 35, 36) and rested upon ancient vaulting, so that he believed it to be the principal building of the villa. Between it and the Via Latina he observed a 1 I noticed some paving stones in the field wall on the N.N.E. side of the modern road diverging just E. of the Casa Santangeli, which may have belonged to it. If so, then Grossi-Gondi marks it a little too far to the W. (Compare also the next footnote.) 2 Cozza-Luzi, // Tuscolano, 95. There are indeed remains of the pavement of an ancient road 2'05 metres wide going in a S.E. direction in the path coming to the Abbey from the Vigna delle Monache, just before reaching the Abbey. 3 The inscriptions, copied at Grottaferrata, without note of their provenance are C.I.L. xiv. 2538, 2544, 2545, 2550, 2551, 2554, 2561a, while ibid. 2438 was found vaguely 'near Grottaferrata,' and so was 2566 (a Christian inscription). 4 Piacentini, Comm. Graecae pronunciationis, 62 (the original authority for the discovery) wrongly describes this as a mark on a lamp—or at least so De Rossi understands him. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 229 pavement in opus spicatum, and other mosaic pavements and many walls had he was informed been found between it and the high-road. It is now the site of a large convent of Franciscan nuns, the construction of which has no doubt obliterated these remains. To the S.W. of the convent is a massive embanking wall with large buttresses, in opus incertum. The other portion, to the S.E., consists of a platform with some vaulted chambers within it (among them a chamber in the vault of which Stevenson noted a terracotta drainage pipe, and a cryptoporticus, in which was found a fragment of a marble ceiling) on the S. angle of which stands the house which was formerly the Casino Beccari (though when he made his notes it belonged to Nicola Santovetti) but is now included in the property of the nuns. Near the house he saw an ancient wall and other fragments ; and to the N.E. of it a reservoir with several chambers, part of which has been converted recently into a cellar, while the rest serves for the foundations of a new building. Close to this, by the entrance gate, is a marble cippus with a portrait of the deceased in the pediment, and a now illegible inscription, which Stevenson also saw. In this vineyard too Cozza copied the unimportant sepulchral inscription C.I.L. xiv. 2533. A sarcophagus lid of peperino in shape like the boiler of a railway engine (cf. Papers, iv. 118) is recorded as having been found about a.mile from Grottaferrata towards Rome, and copied by Stevenson. In his notes (Vat. Lat. 10572, 19, 35) he states, on the authority of d'Ottavi and Teodoro Croci, that it was found in the so-called ' prato di Grottaferrata' in the Carbone property near the Fontanaccio, i.e. in the E. portion of the present Vigna delle Monache. Here were, he notes, also found walls and a large base which Santovetti had seen, and a bas-relief, and a cornice believed to be identical with that over the door of the Abbey church (which is Byzantine). The sarcophagus lid bears an unimportant sepulchral in- scription {C.I.L. xiv. 2555). I saw it in 1907 at the Casino Santovetti, a little to the E. of the Abbey of Grottaferrata. There are no certain traces of antiquity in the path leading S.E. from the former Casino Beccari to the Abbey. We may now mention various discoveries of the seventeenth and eighteenth centuries, at or near Grottaferrata, the sites of which cannot be very clearly fixed. Pococke, writing in 1730, notes (B.M. Addit. MS. 22981, 60) that at 230 THE BRITISH SCHOOL AT ROME.

Grottaferrata ' they lately dug up three marble heads much defaced, one of a young woman, the other I thought was a boy, and the third is young with hair plaited, which might be Cicero's last wife, and the others his son and daughter. There is a very fine bas-relief over a door, but it is broke, one man has the legs of another in his hand and there it is broken and 'tis concluded to be the Roman military charity; here the Hermaphrodite with woman's breasts and man's clothes which I saw in Villa Pamphili near Rome was found.' The last reference is to the Apollo (which went under the name of a Hermaphrodite, cf. Clarac 667, 1548A) described by Matz-Duhn, 188, which is still in the Villa Pamphili {infra, 234) and to the fragment of a relief still at Grottaferrata, published by Winckelmann {Mon. Ined. 136). Gavin Hamilton made excavations at an unknown site near Grotta- ferrata early in 1773. We find him writing to Lord Shelburne on December 26th 1772,' I have made (few ?) discoveries of late, but after Carnival sh(all dig) at Grotto Ferrata, famous for the Villas of (Sulla) and afterwards of Cicero, who have formerly (spoiled) Greece of what they could find excel- lent,' and on January 29th 1773, ' I am now making my excavations near Grotto Ferrata, where I have begun with some success, having already found some very fine basso-relievos (and) which are already bespoke for the Pope. His Holiness seems to have very extensive views with regard to the new Museum, and the difficulties of sending away antiques increase daily.' (A. H. Smith, Catalogue of the Ancient Marbles at Lansdowne House, p. 64.) Exact details of what he found are nowhere given by him. Of the Abbey of Grottaferrata itself we cannot here speak in detail: its mediaeval and modern history will be found in Tomassetti, Via Latina, 142 sqq. and in Rocchi's La Badia di Grottaferrata, ed. ii. (Rome, 1904). See also Lanciani, Wdiiderings in the Roman Campagna (London 1909), 266. I may notice, however, that Frederick 11. removed from the abbey in July, 1242 two bronze statues of a man and of a cow which had long stood there as ornaments of the fountain, and carried them off to Lucera, whence they have long since disappeared (Ughelli, Italia Sacra, x. pp. 238, 289: Pertz, Mon. Germ. SS. xviiii, p. 231). The inscription set up by Cardinal Barberini under one of the bas-reliefs which still remains at Grottaferrata is given by Mattei, Tuscolo, 74 : he speaks of it as tabula velut e naufragio relicta. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 231

Nor shall I attempt here to give a catalogue of the excellent museum of antiquities from the district which it contains, a collection which is con- tinually growing in interest. But I may call attention to the bronze plate, once tied round the neck of a slave, published in C.I.L. xv. 7188 (infra, 267). It may, too, be worth while to remark that the round base in the Villa Pamphili Doria with Antoninus Pius (?), Roma, Ares, etc. (Matz-Duhn, 3684) was drawn in the garden of the Abbey in the seven- teenth century (cf. the Dal Pozzo drawings at Windsor, Bassirilievi, III. 34, 35, 45); and that a fragment of a Greek funeral relief in Palazzo Colonna (Matz-Duhn, 3728) is to be united with the fragment mentioned supra, 230: Braun {Ant. Marrnorwerke, I. Taf. ix a : cf. Bull. Inst. 1838, 22) first recognised that the two pieces belonged to the same relief. An inferior and much injured copy of the group of a cow and a boy in the Sala degli Animali in the Vatican (n° 234) exists in the Museum ; cf. Amelung, Die Sculpturen des Vatikanischen Museums, II. p. 393. Various inscriptions have been copied at the abbey of Grottaferrata, the provenance of which is not known. Besides those already named we must mention C.I.L. xiv. 2519, the cinerary urn of Celadus, C. Caesaris dis- pensator, and ib. 2561% a late fragment from the pavement of the church.1 The Christian inscription ib. 2566 was found near Grottaferrata in 1765 : see Bull. Crist. 1875, tav. viii. fig. 1. C.I.L. xiv. 2535 was found 'in fundo quodam monasterii Cryptoferrat- ensis.' It is a sepulchral inscription with an appeal for the sanctity of the tomb per deos superos inferosque te rogo tie ossuaria velis violare. M. Calpurnius M. L. Sulla Calpurnia M. L. Fausta liberta. E. Q. Visconti, in his MS. preserved at Paris, gives as found in 1780 ' nello scavo di Grottaferrata,' an excavation of which we have unluckily no further details, C.I.L. xiv. 2520 (a sepulchral inscription). Nor do we know exactly where C.I.L. xv. 1030. a. 13 was found ('ai Montiglioni' near Grottaferrata) nor ib. 313. 18 copied byGiorgi'on the road to Grottaferrata in the ruins of an ancient path, on the right.' The abbey rests upon the substructions of a Roman villa. (Whether the opus quadratuin blocks, noted by Nibby, Schede, i. 56 in the wall of the church on the side towards the garden, belonged to this or to 1 I may here call attention to the votive inscription seen in the sixteenth century in the pave- ment of the church, which mentions a bishop Fortunatus (of Labici) of the fifth or sixth century. (De Rossi, Bull. Crist. 1872, 112; Duchesne, Arch. Soc. Rom. Stor. Pair. xv. 1892, 496.) There was no bishop of Tusculum before 1110. 232 THE BRITISH SCHOOL AT ROME. some other ancient building is doubtful.) On the S.W. side overlooking the deep narrow valley of the Acqua Mariana, is a cryptoporticus, originally double ; the walling is faced with opus reticulatum of selce, with quoins of the same material. Several photographs of the details (B 1-7, 63, 64) have been taken by the Ministry of Public Instruction, cf. Catalogo delle Fotografie del Gabinetto Fotografico, Rome, 1904, p. 43. PL XXIV. fig. 2, from a photograph of my own, shows the abbey from the opposite side of the valley of the Marrana Mariana. This villa has by many been identified with the villa "which Cicero owned in the territory of Tusculum, to which there are so many references in his works. The evidence for the determination of the site is carefully examined by Grossi-Gondi (pp. 64 sqq.) who states the other rival views— that of Zuzzeri, Antica villa scoperta sul dorso del Tuscolo, recently adopted by Schmidt, Cicero's Villen (reprinted from Neue Jahrbucher f. d. Klassische Altertum, ii. (1899)) 30 sqq., who places it near the Villa Rufinella {infra, 338); that of Canina {Tuscolo, p. 90), who places it near the tomb of M. Metilius Regulus {infra, 241) but to the east of it, extending from the Ponte delta Macchia to the Villa Muti; and that of Albert, adopted by Lanciani, who identifies it with a villa on the Colle delle Ginestre {infra, 256).1 From Cicero's own writings we learn (1) that the villa of Lucullus was not very distant from his own {De finibus, iii. 3 ; iv. 28 ; Acad.prior ii. 48), (2) that the villa of Gabinius {infra, 251) was also not very far away, (3) that Cicero paid a water rate to the people of Tusculum for the Aqua Crabra, which must therefore have supplied his villa (De lege agraria, iii. 2. 9) ego Tusculanis pro aqua Crabra vectigal pendant, quia mancipio fundurn accept; si a Sulla mihi datus esset, Rulli lege non penderem, It is clear from Frontinus that the springs of the Aqua Crabra were situated higher than those of the Aqua Julia, and were inferior in quality to them (infra, 388) and, as we shall see, they rise in a basin about six hundred metres above sea level.2 It is, however, hardly conceivable

1 Others finally, such as Volpi, Vetus Latium, viii. 87, Zuzzeri, op. cit. 48, and Eschinardi, Esposizione della Carta Cingolana, 374, suppose that Cicero had two villas, and the last named cuts the knot of the controversy by remarking that there were reasons for placing it at Tusculum, and others for placing it at Grottaferrata, that some desired to place it a little way above the Villa Sacchetti (Rufinella), where its ruins might still be seen ; and that one might conclude that there were two for different seasons ! Venuti, in his revised edition of Eschinardi (p. 274) is among those who place it at the Rufinella. 2 Grossi-Gondi (p. 83, n. 2) is wrong in suspecting a misprint in Lanciani's Comentari di Frontino (Mem. Lincei, Ser. III. vol. iv. (1880) p. 321). See infra, 388. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 233 that they supplied the ancient villa near the Villa Rufinella, having regard to the contour of the ground and the course of the modern Acquedotto Aldobrandino (see Canina, Tuscolo, 85, and our maps). But the Aqua Crabra could, there is no doubt, have easily reached any of the other villas ; and, inasmuch as we do not know how much water Cicero took (it is most unlikely that he was the only user of the aqueduct) nor what was the actual course of the ancient aqueduct, we cannot infer much from the present course of the two channels which now receive the springs of the Valle della Molara. The other arguments which have been brought in to determine the site of the villa are (a) various objects which have been or are said to have been discovered on the various sites proposed (a) near the tomb of Metilius Regulus, some inscriptions (in reality either spurious or not belonging to the site : cf. C.I.L. xiv. 222*), two statues, one male and one female, crowned with laurel, a statuette of a boy, two bas-reliefs and two headless busts, one with the name of Cato, the other with that of Cicero, the discovery of which would prove little, being if anything in favour of Canina's view, though the evidence, even if trustworthy, is insufficient1 (Grossi-Gondi 71);

1 The fact that the account of Mattei (Tusculo 72) is untrustworthy as regards the inscriptions need not condemn it as a whole, for we have an independent version of the same discovery in a MS. now in the library of the Episcopal seminary at Frascati (Cod. Tusc. 14, i. II f. 188), which is given by Lanciani, Bull. Com. 1884, 190. From this we learn that the site belonged to one Luigi Ceppi, that Cardinal Francesco Barberini had the statue of the woman, the two busts, and a fragmentary group, and that Cardinal Massimi had (Mattei says bought) the male statue, the statuette of a boy and the two bas-reliefs, which measured six palms (1 "33 metre) square each : they were placed in his palace in Rome at the Quattro Fontane, and sold on his death (when the palace was also sold) by his brother, and removed to France. The fragmentary group which Cardinal Barberini had is thus described, ' a very beautiful broken fragment, on which one sees two small feet without legs, and two other feet with the thighs, with a cloak over the thigh of one of the boys (the group cannot represent anything but two boys embracing) and these fragments are now in front of the Palazzo Barberini, where the sculptors are at work and where there is a large quantity of various ancient fragments found partly at Grottaferrata (cf. the inventory published in Documenti Inediti, iv. pp. 56 sqq., Nos. 47, 53, 63) and partly in the plain now called Le Frattocchie' [infra, 282). The Cardinal Massimi meant is no doubt Cardinal Camillo, whose collection of antiquities in his palace at the Quattro Fontane, is spoken of in the Nota delli Musei (p. 33) placed at the end of the 1664 edition of Lunadoro's Relazione della Corte di Roma: he was made Cardinal in 1670 and died in 1679. Grossi-Gondi is probably right in supposing that the Cod. Tusc. is mistaken, and that the Cardinal Barberini meant is really Cardinal Carlo {infra, 253). None of the antiques can now be traced, though the two busts are mentioned in the inventory above cited, which dates from 1738, in the list of fragmentary statues, etc. (p. 59, No. 87, 'two square bases like terminal figures without head and arms, one of Marcus Cato and the other of Marcus Tullius Cicero, one palm (0-22 metre) high and wide excavated at Grottaferrata'), while the group of two boys might correspond with several of the fragments described (e.g. p. 70, No. 287). The history of the Barberini collection, like that of all the great Roman collections of sculpture, has yet to be written. 234 THE BRITISH SCHOOL AT ROME.

(J8) at the Abbey of Grottaferrata—the circular base mentioned supra, 231 which was wrongly believed to be the Tpaire^o<}>6po

I have made an attempt to deal with the collection formed by Cardinal Ippolito d'Este in his villa at Tivoli in Archaeologia lxi. 219 sqq. 1 It was dedicated to Giambattista, the son and successor of Camillo Pamfili; and the imprimatur was given by Fr. Hyacinthus Libellus (1660-1668). CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 235

Grottaferrata—a tradition, however, which, as Grossi-Gondi well points out (pp. 75 sqq.), is not traceable earlier than the middle of the 15th century: Pius II. in his Commentaria, speaking of a visit of May 30th, 1463, says monasterium est in agro Tusculano situm, Marianum inter et Lucullanum, quo in loco Ciceronis villam fnisse putant et ibi quaestiones Tusculanas editas. But in the chronicle of Petrus Aurelius, Bishop of Sinigaglia, who described the journey of Gregory XI. in 1377 from Rome to Anagni, we find no allusion to this tradition in the description of the Abbey; and it probably grew up in the time of Cardinal Bessarion, who became commendatory Abbot of Grottaferrata in 1462, and whose relations with the humanists of his time are well known to us. Even if it were proved that the site was continuously inhabited, and especially from the 5th to the nth century A.D., when S. Nilus founded the Abbey (Grossi-Gondi, p. 80, n. 5) this would have no bearing on the question. (c) the comment of a scholiast on Horace, Epod. i. 29, neque ut superni villa candens Tusculi, which runs thus :. Tusculi superni: hoc est in monte siti, ad cuius latera superiora^ Cicero suam villain habebat Tusculanam- This passage is made use of by Zuzzeri, but it furnishes an argument of little value ; nor does the description of the villa of Gabinius as ad hunc Tusculani [in monte] montem in Cicero's speech in Pisonem (21. 48) give us any sufficient ground for adopting this view, and the reading is quite uncertain. That which I have given is adopted by Muller ; but some MSS. give simply in hunc Tusculanum montem. In any case mons is much too vague to compel us to place the villa actually on the hill of Tusculum. Lanciani tends, as I have said, to adopt the opinion of Albert {Bull, Com. 1884, 192) and Grossi-Gondi (pp. 92 sqq.) follows him, while admitting that there is no certainty to be attained, in thinking that the Colle delle Ginestre is the site which corresponds best to the indications which Cicero gives us. Unfortunately (for it would be far more interesting if it were possible to come to a positive conclusion) I am obliged to say that I think we must, in default of further evidence, refuse to attempt to identify the site more precisely, and that the one really certain indication is that given us by Cicero's mention of the Aqua Crabra {infra, 388). Grossi-Gondi brings forward two other arguments from passages in 236 THE BRITISH SCHOOL AT ROME.

Cicero's own letters in support of Albert's view, which we must examine carefully before dismissing the subject; but I cannot, as I have said, admit their validity. I do not think that Grossi-Gondi is right in the sense that he gives to the passage, a part of which he quotes, from Cicero's letter Ad Atticum xii. 36. Fanum fieri volo, neque hoc mihi eripi potest. sepulcri similitudinem effugere non tarn propter poenam legis studeo quain ut maxime adsequar a-rrodecocnv : quod poteram, si in ipsa villa facer em, sed, ut saepe locuti sumus, commutationes dominorum reformido. In agro ubicumque fecero, mihi videor adsequi posse, utposteritas habeat religionem. He maintains (p. 94) that had the ager selected been bounded by a .public road, the inconvenience of a change of proprietors could have been avoided by erecting the monument, according to the then prevailing custom, on the edge of the road, and declaring on the monument itself how much space infronte et in agro was allotted to the sacred area of the tomb. He goes on to argue that Cicero could certainly have erected it even in the grounds of his own villa, had this been situated upon an important road, upon which it was the custom to erect monuments: instead of which, Cicero insistently asks Atticus to find him another site distinct from his Tusculan villa. The latter, therefore, he maintains, was not touched by any main road ; and he uses this argument as an important element in the deter- mination of the site of the villa. I must confess that the first few words (which Grossi-Gondi omits) fanum fieri volo ... sepulcri similitudinem effugere studeo seem to me not at all consonant with the idea that Cicero could have contemplated erecting the fanum along a public thoroughfare, where, as we know (from examples, it is true, of the imperial period), the external form of a temple was so commonly adopted for a tomb. Instead of this Cicero is always begging Atticus to buy him a garden, and suggesting to him various proprietors (Cic. Ad Att. xii. passim) ; and it appears to me to be clear that his desire is to erect it in a fairly secluded spot and yet not entirely out of the way: cf. Ad Att. xii. 19, where, after expressing his fears that his property at Astura, though suitable, might too often change hands, he adds, cogito interdum trans Tiberim hortos aliquos parare et quideni ob hanc causam maxime: nihil enim video quod tain celebre esse possit. (I can hardly agree with the words I have italicised in Tyrrell.and Purser's note ' Cicero was desirous that the shrine dedicated to his daughter should be in a central site, where the traffic would be constant and abundant' for it will be CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 237 noted that he speaks here too of his desire for a garden, not of a site on a high-road). I am therefore inclined to interpret in agro more simply. Cicero is afraid that if he erects the monument in the villa itself, close to the house, a subsequent owner may find it in his way, and remove or alter it: on the other hand, if he erects it away from a house, or in grounds specially set aside for the purpose, he thinks that, wherever the chosen site may be, he will be able to secure the respect of posterity. The other passage cited by Grossi-Gondi (p. 92) again only in part (he omits the second sentence) ego in Tusculanum fiikil sane hoc tempore: devium est TOU airavraicnv et habet alia Bvcrxpva"ra- Sed de Fonniano Tarracinam pridie Kal. Ian.; inde Pomptinam summam, inde Albanum Pompeii, ita adurbem III. Nonas, natali ineo {Ad Att. vii. 5) must similarly be taken in a wide sense. It does not mean that the villa was a mile or half a mile, more or less, from the Via Latina, or, as he says (p. 96), of the Colle delle Ginestre, ' distant from an immediate or easy approach from the Latina' (except that this is a rather steep hill, it is as close to the road as it well could be); but that the Via Latina was not one of the main highways of Italy, like the Via Appia, on which are situated all the places he names in his letter, and along which he himself was travelling from Brundusium, where he had arrived on November 25th (50 B.C.) on his return from Cilicia {Ad Att. vii. 2). Tyrrell and Purser rightly translate 'it is out of the way for chance rencontres' (with travellers of his acquaintance who would convey his letters). The next letter (vii. 3) was written on December 9th from Pontius' villa in the territory of Trebula, a town which probably stood on the hill of Tripaola, and had a post station (vicus Novanensis or ad Novas) on the Via Appia between Calatia and Caudium (see Nissen, Italische Landeskunde, ii. 753, 810). Where the next few letters were written we do not exactly know. The one we are examining (vii. 5) says sororem tuam non venisse in Arcanum miror. This was an estate of Cicero's brother Quintus, the exact situation of which is not certain. Nissen {pp. cit. ii. 673) refers arx Fregellana in Liv. ix. 28 (cf. Diod. xix. 101) not to the citadel of the town of Fregellae (as does Colasanti, Fregellae, 139) but to the hill fortress of Rocca d'Arce, 504 metres above sea-level, above the modern village of Arce, five miles N.N.E. of the site of Fregellae, and defended on the most accessible side, he says, by a polygonal wall. This seems to me very reasonable : for the name Arx. was given to the 238 THE BRITISH SCHOOL AT ROME. place by the Geographer of Ravenna (iv. 33) and Paulus Diaconus {Hist. Lang. vi. 27), belonged to it through the Middle Ages (when it was regarded as impregnable) and still clings to it. Moreover, it supplies a good derivation for the name Arcanum, which we cannot otherwise explain. Hiilsen, indeed, (s.v. Arx in Pauly-Wissowa, Realencydopddie, ii. 1493) regards it as certain. Cf. Mommsen in C.I.L. x. p. 555. It had, however, probably in Cicero's time already been incorporated in the territory of Arpinum (Ad Att. i. 6..2). From what follows I think it is clear that we must not suppose that Cicero himself had actually gone so far out of his way as this : he had merely heard the news that Pomponia (Atticus' sister and his own brother's wife) had not come there. The next letter (vii. 6) contains no local indications, and the next (vii. 7) intimated that he would be a day later in reaching Pompey's Alban villa, and consequently also in arriving in Rome. The next, however (vii. 8), speaks of his arrival at Formiae on Dec. 26th, and in the next (vii. 9) we have the explanation of the phrase we have been examining: ' Cotidiene' inquis ' a te accipiendae litterae sunt ?' Si habebo cui dem, cotidie. ' At iam ipse ades! Turn igitur, cum venero, desinam. Unas video mihi a te non esse redditas, quas L. Quinctius, familiaris metis, cum ferret, ad bustum Basili1 vulneratus et spoliatus est, and on the fourth of January he was already in Rome, where he laid down his imperium. The reason for his preferring the frequented route along the Via Appia was thus obviously that it afforded better opportunities of sending and receiving letters, and especially for his correspondence with Atticus, who was in Rome, and to whom he was writing constantly, expecting to meet him either in Pompey's villa or in Rome (Ad Att. vii. 8. animadverteram posse pro re nata te non incommode ad me in Albanum venire III Nonas Iannar. Sed, amabo te, nihil incommodo valetudinis feceris. Quid enim est tantum in uno aut altero die ?). So that we cannot out of this passage, any more than out of the first, draw any indications for the site of his Tusculan villa.

1 Cf. Asconius In Milon (p. 50, Orell.) Via Appia est prope urbem monumentum Basili, qiii locus latrociniis fitit perinfamis. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 239

XIV.—THE VIA LATINA FROM THE TWELFTH TO THE THIRTEENTH MILE.

Just after the twelfth mile1 of the ancient road, which falls approxi- mately at the modern tramway junction (where the line to Frascati diverges from that to Grottaferrata) a road branches off through the Macchia di Grottaferrata, almost due N., which probably follows the line of an ancient road.2 I noticed in 1904 fresh paving stones in the fieldwall, probably found in making the vineyard on the E. A little way along it, on the W., Stevenson noted in September, 1891, in the vineyard of Teodoro Croci, the pavement of the Via Latina, which is thus marked a trifle too far south in my map. Here too in 1854 were found many blocks of' sperone,' a kind of tufa {Lapis Gabinus) belonging to the substructions of the road (Atti. Mitt. Lav. Pubbl.^2%7, cited by Tomassetti, p. 141 n.). Just to the N., in the vineyard of Antonuccio Vendetti, a lead pipe was found ( Vat. Lat. 10572, iv, cf. ibid. 36). I cannot fix the exact locality of the following discovery referred to in Stevenson's notes. A letter of April 4th .... from Pasquale Antini {cod. cit. 20 v) informed Stevenson that he had found in the Santovetti property a rectangular shaft about 1 metre by O'5O metre and about 3 metres deep, and other channels of peperino of various sizes, one about C30 metre in diameter, and various points where the ancient road existed (the reference is in all probability to the Via Latina). Stevenson {cod. cit. 23T) appears to have then visited the place itself, for he noted that the ancient road passed near the shaft, and that here was the vineyard of Costantino Longacci. In Bull. Com. 1902, 109 Grossi-Gondi describes the discovery in the Vigna Tappi (formerly Passamonti)3 near the tomb of Metilius Regulus of the pavement of the road : its direction, he says, confirmed Rocchi's theories. Remains of other tombs were found and near one of them an unimportant inscription of one Fabius Augustalis, and other objects. 1 Near the twelfth mile was a church mentioned in the Bull of Sergius I, and the Regestum of Gregory II: basilica S. Petri intra massam Marulis via Latina milliario ab urbe plus minus XII. (Armellini, Chiese di Roma, 890: cf. supra, 224), 2 This is also the opinion of Nibby (infra, 240) and Stevenson. 3 It is mentioned as existing in this vineyard by Cozza in Giornale Arcadico cxc. 115=/^ Tuscolano, 123. Stevenson ( Vat. Lat. 10572, 19) notes that he had been informed of the discovery of two marble sarcophagi in this vineyard. 24Q THE BRITISH SCHOOL AT ROME.

The ruins marked to the W. of the road in my map, were brought to light in the course of the tramway works : a vaulted substruction in concrete, 2"5O metres wide, was found, and to the E. of the road a very large reservoir in opus reticulatum was discovered, consisting of three parallel vaulted chambers each 53 metres in length and 3*41 to 3^2 metres in width, divided by walls 1^50 metre thick: there were several apertures in these walls, each 1-65 metre in span. A marble cinerary urn, without inscription, and some pottery and glass were also found {Not. Scavi, 1904, 273) ; and I also saw a late tile burial in the tramway cutting. Further N., on the west side of the road, is a large and prominent tomb, a lofty square mass of selce concrete, with a (probably) modern chamber in the upper portion. Grossi-Gondi gives a photograph of it (tav. x.). It is very possible (though not, I think, at all certain) that C.I.L. xiv. 2501 may have belonged to it. One fragment was copied in the territory of Grottaferrata in 1673, another built into the Casale Santangeli, some 500 yards to the S.W., and a third is said to have been excavated near that casale. It is the sepulchral inscription of M. Metilius Regulus, consul ordinarius in 157 A.D. {Prosopographia. ii. p. 371. No. 385.) In Nibby's time (see below) the vineyard belonged to Gaspare Baccari, having previously been the property of Silvestro Tiberi. He notes the existence close to it of fragments of columns and Corinthian capitals in peperino, and fragments of marble. I have seen tufa columns and similar fragments myself. To the W. of this tomb the tramway works brought to light (and destroyed) the remains of a structure in opus quadratum of tufa, the blocks being 1 metre thick {Not. Scavi, loc. cit.). Grossi-Gondi notes it as interesting, but does not give full particulars, speaking of it as though it were the stylobate of a tomb or a temple. Nibby {Sckede i. 71-73, no) in a detailed description of the ruins in this district (which he visited in October, 1822), which does not appear in his published works, speaks of it as having a facade facing S.E., in opus quadratutn of tufa, which appeared to have been added later. From this one descended into a small corridor of selce concrete, and thence into a chamber faced with small pieces of incrustations from the Anio (Italian writers call them ' Tartari tiburtini') as a method of decoration. The door was central neither with the passage, nor with the chamber, which was a nymphaeum or bath: a kind of step CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 241 had been added later, which partly covered the facing of ' tartan.' A water channel or specus, of the height of a man, the direction of which was almost diagonal to the entrance, seemed also to be an addition. It is apparently to this place that the erased plan in Stevenson's notes (Vat. Lat. 10572, 25) refers. To the E. of the so-called tomb of Metilius Regulus and of the modern road, a path runs E., which Nibby marks in his map as following the line of an ancient road discovered and destroyed not long before his visit. To the N. of this, in the then Vigne Zocchi, Vannelli, dell' Osso, and Amadei, he saw the remains of a large villa in opus reticulatum of selce, with quoins of the same material, facing and parallel to the road through the macchia di Grottaferrata. On the E. portion of the site was a large rectangular open space, probably a piscina or fountain basin, with an apse facing E., measuring some 130 metres in circuit. He noted the existence of marbles of all kinds and of fine architectural fragments as indicating the magnificence of the villa. To the S.W. of the piscina he saw remains of chambers in brickwork of the first century A.D. : one of these was decorated with niches in which (as in the domus Augustana on the Palatine, in the portion under the Villa Mills) the door of communication was in the niche itself. The other walls were mainly long substruction walls, one on the front of the villa having square niches. He mentions that he saw a column of grey marble, and architect- ural fragments in peperino. The modern house (marked 344 on the map) is built, of course, of ancient materials. Stevenson (Vat. Lat. 10572, 23) has some notes as to this villa. A large mosaic and two trapezophori were reported to have been found there ; and on the terrace where the house stands a floor of opus signinum was discovered in which were various round pits O'So metre in diameter and the same in depth, with marble at the bottom. He also noticed column drums of grey marble and peperino, a small head of a boy crowned with ivy, etc., and also the ancient road, some 17 yards from the edge of the macchia. He mentions too (ibid. 341, 57V) a bas-relief with putti in the main street of Grottaferrata, which was found here or hereabouts, in the Quarto Cipriana. This name, which belongs to the locality (from the Via Latina as far as the S. boundary of the macchia), gives Nibby a reason for calling this the villa of Cato, inasmuch as he made Cyprus a part of the Roman Empire : but for this view there is but little foundation. To this villa would belong, 242 THE BRITISH SCHOOL AT ROME. as Grossi-Gondi points out, the discoveries upon which Canina relied {supra, 233) for determining the site of the Villa of Cicero; but the evidence derivable from them is too slender. There is hardly more reason for supposing this to have been the villa of the Vibii, as Grossi-Gondi (p. 99) is inclined to do. The inscriptions which he cites as found here are none of them other than sepulchral, and are of people of no great position, so that the evidence for his theory must be considered insufficient. De Rossi {Ann. Inst. 1873, 190) only speaks of tombs of the Vibii in this locality. The inscriptions in question are the following :—(a) the triplicate inscriptions1 C.I.L. xiv. 2556, 2557, 2558—the first (entire) found below the Vigna Cavalletti (cf. Papers, iv. 127) in 1857, the second (a fragment) between Borghetto and Ciampino in 1873, and the third (again a fragment) in the Vigna Gambini at Campovecchio {infra, 269) relating to the concession of a sacred area by one Varena Sabina, in which M. Publilius Strato, freedman of Publilia and of C. Vibius Rufus, erected a tomb for himself, Varena Sabina, and members of her family, and for some of his own relations, freedmen and freedwomen, {b) a sarcophagus bearing the sepulchral inscription C.I.L. xiv. 2525 (Matz-Duhn, ii. 2572), (V) the cippus bearing the fragmentary sepulchral inscription C.I.L. xiv. 2560.2 The old path through the macchia of Grottaferrata soon turns N.N.W., and is joined by the path coming from the cemetery of Grotta- ferrata. To the W., just outside and north of the macchia, Stevenson in his map has marked the pavement of a road, and in his notes {cod. cit. 547) has a long passage in regard to it. He saw a few paving stones on the edge of an oliveyard and the macchia (which are still there) and was informed by the sacristan of the church of S. Pietro that the pavement had been found a little further E. in his vineyard, and also, along the road, buildings, tombs 'a cappuccina' with lamps and a ring with a cornelian bearing a Gnostic device or inscription, a fragment of a lead pipe, with five or six letters, which had been sold to Monsignor Battamelia, etc.—also a rectangular shaft with footholes 3 metres deep, leading to a drain parallel to the road. The road must have come out, he thinks, near the entrance to the Villa Muti from this side : but there are no pavingstones in the S.W. 1 I omitted to mention the third copy in Papers, iv. 127. 2 Here is (or was in the 'eighties) the Vigna Consoli, and here was copied the unimportant sepulchral inscription C.I.L. xiv. 2434. In the Quarto Cipriana another inscription of this nature was found in 1894, and is now preserved at the abbey of Grottaferrata [Not. Scavi, 1894, 313). CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 243 enclosure wall of the Villa Muti, which is entirely covered with cement, so that it is impossible to say whether the road went on eastwards towards the Villa Montalto beyond its junction with the road through the macchia. Going westwards, he notes that it was found lower down the hill in the Vigna Muti: there is a house there (not marked in the map) with debris of a villa, but no certain paving stones. In this vineyard, or at point 344, were found the brickstamps C.I.L. xv. 213, 686, 708 a, 1762, 1770, 1798, 1802, 1861, copied by Giorgi on Oct. 10, 1732, and Feb. 23, 1733, in the Vigna Amadei, beyond the Villa Rocci (now Muti), and ibid. 361, 12 (123- 125 A.D.) copied in the same vineyard by Lupi. Giorgi also saw in a vineyard house below the Villa Spada and opposite to the Villa Amadei a sarcophagus with the inscription C.I.L. xiv. 2699. The Villa Muti itself (formerly Villa Varese and Villa Rocci) occupies the site of a large villa, of which, however, few remains are now visible. The account of Cod. Tusc. 14. i. 11, f. I4iv sqq. is a mere translation of Kircher, Latium, 73 sqq. It is wrongly referred by Lanciani {Bull. Com. 1884, 201) to this site: the plan on p. 73 is that of the villa at Fontana Piscaro {Papers, iv. 135) and that of the round piscina (p. 74) is that of the circular reservoir to the N.W. of that villa {Papers, ibid?)} Nibby in Analisi, iii. 354 (cf. Schede, i. 75, where he deduces its shape only from the appearance of the ground, the rest having, he says, perished) makes the same error.2 He, in his description, notes the air of desolation which then pervaded the villa, and gives copies of C.I.L. xiv. 2605, 2721/2 {infra, 249), both first recorded here by Fabretti, with no information as to their provenance. He also notes the existence, under the inclined plane leading to the upper terrace, of some walling in polygonal work or selce, with a later facing of opus reticulaturn of the same material, of which in Schede cit. he gives a sketch. In Not. Scavi, 1884, 157, Lanciani describes, with a sketch plan, several rooms in opus reticulaturn and opus quadratum with fine mosaic pavements (belonging thus probably to the first century A.D. and repaired in the second century) found on the uppermost terrace : apparently the angle of an atrium or peristyle was found, with Doric columns of peperino. Three brickstamps {C.I.L. xv. 272-123 A.D. : 1333—first century A.D. (?) : 1 PI. XXXV. Fig. 1 shows the facing of the platform W. of this reservoir. 2 He remarks, too, correctly that this site belonged once to the Rocci, then to the Varesi, and finally, after being bought by Cardinal York for the episcopal seminary, took the name of Vigna del Seminario. R 2 244 THE BRITISH SCHOOL AT ROME.

1986-Hadrian (?)) were found, also some fragments of terracotta friezes, and the hand of a discobolus, finely worked. The substructions in opus reticulatum, according to Lanciani {Bull, cit.) extend for a length of over 150 metres. The inscription C.I.L. vi. 1625 a, a dedication to M. Petronius Ho- noratus {Prosopographia, iii. 207) was copied here by Doni early in the seventeenth century. At this time a small house, erected in 1579, was trans- formed into a fine palace by Monsignore Arrigoni (d. 1616). It is, therefore, considered probable by Lanciani, and by Grossi-Gondi (op. cit. 121), that the inscription was found here, and not brought from Rome : so that the latter, with some reserve, attributes the villa to this personage. Here was also copied the sepulchral inscription of one Claudius Verus, an evocatus {C.I.L. xiv. 2617). The walling seen by Nibby and Lanciani is now almost entirely hidden. There are, however, in the garden various antiquities—'numerous statues of no great merit, though one, in the upper garden, is of interest, if only because there is a coloured drawing of it among the drawings of the Dal Pozzo-Albani collection, formerly in the possession of Sir A. W. Franks, and now in the British Museum (vol. i. f. 143). It represents a Roman lady with a head-dress of the Flavian period lying on a couch with her head on the pillow: she rests on her left shoulder, holds a garland in her left hand, and her drapery in her right : the statue is 177 metre long and O'66 metre across: there is also a good piece of a frieze with bucrania, 177 metre long, C85 metre high, which Nibby also mentions. Among the antiquities in the Villa Muti many were probably found in Rome, and brought to it for its adornment. Among these we may reckon the inscription Kaibel I.G. xiv. 1110 relating to the Synodus Heradeistarum, which was established near the baths of Trajan. The tramway from the Villa Muti follows a new road across the valley to Frascati ; but the older path descends due N. There are no certain traces of antiquity in it: at a chapel of the Crocefisso it joins the path from Torre di Micara to Frascati (see Papers, iv. 133). To the W. of this point Stevenson noted unimportant traces of antiquity at two places where now nothing is visible {cod. cit. 44) and a pilaster capital at the house of the Vigna Senni on the north of the path, which is still there, as are also other marble fragments. Going further W. we pass the path from CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 245 the Villa Montioni, and then reach the entrance gate of the Vigna Bevilacqua. Outside this there are, as Stevenson notes, four paving stones in situ : the direction of the road to which they belonged is uncertain. Stevenson was told by Micara that he had found the ancient road between the entrance gate and the Torrione di Micara, and that it then ran across the modern path to the oliveyard on the other side (by T on our map) and came out at this point; but he suspected the information. If this were true, we might probably suppose that a branch of it would have gone on in the same direction to the point W. of the Villa Muti, where two (or even three, see p. 242) probable lines of ancient road already join, at the N. edge of the Macchia di Grottaferrata ; but along the diagonal path running S.E. to this point there are no certain traces of antiquity. I saw on its S.W. side a channel in opus incertum 0-59 metre wide and about r2O metre high, roofed with a large block of peperino. Entering the oliveyard, we first reach a small casale, on the edge of the villa as marked in the map, which rests upon an ancient reservoir, in four compartments : in the field-walls Stevenson noticed fragments of mosaic and a broken rectangular brickstamp with large raised letters IC [. Near the larger casale is much debris : lower down is the terrace wall of the villa, partly of polygonal work in a poor and not very compact kind of selce, and partly of opus incertum. Nibby describes it in Analisi, iii. 354 ; in Schede, 1. j6 he states his belief (correctly, I think) that the opus mcertum, which is larger than usual, is a repair, made out of broken polygonal blocks. Stevenson, on the other hand, considers it probable that they are contemporary. Nibby gives a plan ibid. 108. The total length of the platform is about 200 yards, and of the polygonal work, about 100 yards. There is a drain in it C43 metre wide and about 1 metre high, just at a vertical junction shown in the plan and in the photograph. The interior of it is lined with opus incertum ; and in any case there is no doubt that the polygonal work belongs to the villa, and not to any previous construction (for parallel examples see pp. 368, 403). A view of it was made by Dodwell, Pelasgic Remains in Greece and Italy PI. 121 and a photograph is given on PL XXV. Fig. 1. To the N.W. Stevenson noted {cod. cit. 45) the existence of a water reservoir with a single chamber, which I found from his indications : it is faced inside with opus reticulatum of selce, and sunk below ground 246 THE BRITISH SCHOOL AT ROME. level ; it is 370 metres wide and over 18 long—perhaps originally as much as 26 metres. Further N.N.E. is the Casale Piccolomini, where there are some selce reticulatum cubes in the walling, but nothing ancient in situ. On the summit of the ridge above it to the S.S.E., opposite the Casale Bevilacqua, and not far from the path from theTorrione di Micara to Frascati, there is a villa in selce concrete with a reservoir in its platform crowned by a mediaeval or modern house; and to the E.N.E. on the next ridge to the E. (that which runs N.N.W. from the villa Muti) to the S.S.E. of the knoll marked 241 in the map, there is another large villa, with extensive substruc- tures : at one point above the vaulting I saw a herring-bone pavement and the base and the beginning of the drum of a column, 0^42 in diameter, cut out of one block of peperino, still in situ. Higher up the hill are the remains of an enormous reservoir, the concrete of the exterior of which is quite rough. A plan is k here given. » The Villa Pallavicini or Bel Poggio1 very probably occupies an ancient site, but there is no absolute certainty to be attained. A mass of con- ...,.::;,.;.... :.,„ Crete in the bend of the drive ascending on the N.W. side is almost certainly ancient; but the two parallel passages under the upper garden, lined with big roughly parallel- epipedal blocks, show no decided traces of antiquity, and the substruction walls of the garden are, as far as can be seen, quite modern. If, however, the site is not really ancient, the lofty platform on which the garden stands, gives a good idea of what those of ancient villas must have been. There are in the garden two or three possibly ancient marble heads of no merit. The deep cutting through which the path passes S.E. of the Villa Pallavicini may be of ancient origin, but there is no certain evidence. To the S.E. of this cutting, in an oliveyard S. of the Villa Conti, is some ancient debris, which probably marks the site of a villa. Below this the cutting for the new road and the tramway has brought to light two drains cut in the rock.

1 Lanciani (Storia degli Scavi, iii. 56) tells us that its construction was attributed to the Strozzi: it then passed to the Dukes of Ceri (Cesi), then by marriage to the Borromeo and successively to the Visconti and Pallavicini families. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 247

Between the road leading round to the Hotel Frascati and that which leads to the Villa Pallavicini, in the oliveyard are pozzolana pits, several of the galleries of which cut through a round topped Roman drain C45 metre wide, cut in the soil, going about north and south. The Villa Conti (now Torlonia) also occupies the site of an ancient villa. (See Wells, Alban Hills, i. Frascati, 134 for further details.) The substruction wall above the modern highroad has semicircular niches, and the plaster facing of these imitates opus reticulatum: it is not unlikely that this masks ancient work, and there is certainly a substruction wall of opus reticulatum with quoins of selce under the garden-terrace in front of the villa, at its W. angle, and an ancient drain may be seen still lower down. There is also much debris further back, in the ilex grove behind the modern villa, and, in the level space in front of the waterfall, which is on the same level, the outline of a large rectangular chamber may be seen. Above the waterfall, however, there are no traces of ancient buildings. There is, too, further E., not far from the chapel of S. Antonio on the road from Frascati to Marino, a building in opus reticulatum of selce, only part of which was excavated in the course of the enlargement of a pozzolana quarry. A chamber some 830 by 4-10 metres was found, with the south side (one of the two long sides) open : here were two square pilasters, which supported the roof, the bases of which, of peperino, were still in situ : the capitals, of the same material, had Ionic volutes and garlands of flowers. On the right was a narrow space (probably merely an air-space between two walls) 070 metre wide, and beyond it the walls continued. Two brick- stamps were found there, C.I.L. xv. 91 ia (first century A.D.) and a fragment of a lunate stamp OPV • D [ Q and some fragments of terra- cottas. A full account is given by Stevenson {cod. cit. 135), who visited the villa in August, 1892 :x cf. Cronachetta Armellini, 1892, 178, in which further discoveries are mentioned, notably of a wall 28'8o metres long, of a coin of Domitian, and*of a vase representing a chariot race with an inscription which is possibly to be associated with the Sodales lusus iuvenalis {C.I.L. xiv. 2640, infra, 362). Further associations with games are connected with this site, for in November, 1896, the wall dividing the

1 Ten years earlier, in October, 1882, Dressel copied in this villa the brickstamps C.I.L., xv. 388. 5 (Vespasian) 869.3 (en(l °f first or beginning of second century) 2231. a. I (middle of first century) ; but we know nothing of the circumstances of their discovery. 248 THE BRITISH SCHOOL AT ROME. garden from the ilex grove above (which is on a level with the modern villa) fell down, owing to the rain, for a length of some 15 metres : some walling in opus reticulatum came to light under the wall, partly at right angles and partly parallel to it ; and in the soil, which appeared to have been turned over previously, were found various fragments of marbles and a small lead tessera with a figure of Diana running to the left, and the legend subcuriator). (Rostowzew, Tesserarum Plumbearum Sylloge, 863). See Not. Scavi, 1897, 419; 1900, 268. Some earlier discoveries are less exactly noted : thus Mattei, Tuscolo, 64, speaks of no fewer than 18 rectangular rooms with vaulted roofs, in opus reticulatum, 15 palms (3-35 metres) high, 14 (3-13) wide, and 18 (3*97) long, divided by pilasters j\ palms (r66) thick. These must have belonged to the substructions of the villa, and were apparently situated under the ilex grove (for Volpi, Vetus Latium, viii. 117, speaks of the trees growing upon the top of the vaults) which is level with the platform on which the modern building stands, though Montfaucon believed them to be tabernae of the ancient city of Tusculum, while Ficoroni in his commentary on this work {Osservazioni, 15) believed them to be baths. Volpi also speaks of them and of a marble table standing on legs not its own, oval in shape, on the sides of which were winged cupids, animals drawing cars, girls sleeping, etc. The description sounds like that of the lid of a sarcophagus. See also Wells, loc. cit. Turning to the history of the villa, we find that Annibale Caro bought in 1563, from the Abbey of Grottaferrata, a villa at Frascati to which he gave the name of Caravilla (Grossi-Gondi, op. cit. 114). If Ottaviano Caro, who offered in February 1576 some statues to the Commune of Rome, was a brother or a relation of his, we may perhaps suppose that these were found in this villa. Indeed a letter of Annibale of September 14th, 1565, quoted by Grossi-Gondi, op. cit. 116, mentions the villa as being ' nel loco proprio di Lucullo che cosi mi hanno chiarito li vestigi degli grandi monumenti, e di alcune lettere chevi ho trovato.' *The discovery Lanciani supposes to be that of the lead pipes bearing the name of Lucullus, already recorded by him (on the faith of Kircher, Vetus Latium, 73)1 in

1 Kircher says ' altera villa fuit eo in loco, ubi modo hortus Ludovisiorum (the later Villa Conti-Torlonia) est, uti ex inscriptionibus quorundum lapidum ibidem inventorum hisce verbis : L. LUCUL. LUC. F.' but Lanciani, perhaps rightly, supposes the reference to be to a water-pipe. Dessau, on the other hand, treats it as a mere invention, C.I.L. xiv. 209.* CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 249

Comentari di Frontino (Atti dei Lincei, Ser. III. vol. iv. (1880) p. 500, no. 580.) This would thus, supposing Kircher's notice as to the find-spot to be correct, fix the villa of Lucullus on the site of the present Villa Torlonia.1 In that case he cannot have been the founder of the immense villa upon which the town of Frascati is built, and which, being separated from the Villa Torlonia by the ancient road, cannot have belonged to the same owner, as Lanciani had previously supposed {Bull. Com. 1884, 182 : but see infra, 302). From Annibale Caro the property passed to Cardinal Tolomeo Galli,2 and then was sold on his death in 1607 to the Borghese family: in 1613, however, it passed to the Altemps, in 1622 to the Ludovisi,3 and then successively to the Poli-Conti, Sforza Cesarini, and Torlonia families.. It is of course not necessary, as Grossi-Gondi rightly points out (p. 117), to suppose, because Frontinus (Be Aqnis, 5, 8, 10) tells us that the springs of the Aquae Appia, Tepula, and Virgo were situated in agro Lucullano, that Lucullus' possessions extended uninterruptedly over the whole area between these points. But even he wishes to believe that the property of Lucullus extended as far as the Torre di Micara (Papers, iv. 134) and that this was the tomb of Lucullus, which was, we know, in the territory of Tusculum (Plutarch, Lucullus, 43, who tells us that the people wished to bury him in the Campus Martius, but that his brother persuaded them to allow the body to be placed in the tomb which had been prepared for it). There are no other points in the various descriptions of the villa of Lucullus, nor in the classical allusions to it, that would help us to localize it more closely, inasmuch as we are unable to fix the site of the villa of Cicero, near which it was. We know from Cicero's De Legibus (vii. 13) that Lucullus' next neighbour above him was a Roman knight (not as Grossi-Gondi says, Gabinius), while below him lived a freedman ; but this does not help us to determine its position more closely; while

1 A slight additional argument is found by Grossi-Gondi in the fact that C.I.I,, xiv. 2721/2, (supra, 243) is a dedication by two freedmen of the gens Licinia to their patron, and of course might easily have been found here. I may add that Nibby, like Fabretti, saw the fasces and the axe on the left, and adds the word FECIT at the end of the second line on the right. 2 According to documents quoted by Schreiber, Villa Ludovisi, p. 5, he was known as Cardinalis Comensis, though Mas-Latrie makes him a Neapolitan, and from him the villa acquired the name of Villa Comensis. 3 With the Villa they acquired fifteen statues and nineteen heads, an inventory of which is preserved, and is given by Schreiber, op. cit. p. 26. As to their provenance we of course know nothing. 250 THE BRITISH SCHOOL AT ROME. from Plutarch (Lucu/l. 39) we merely learn that he had near Tusculum ' country-houses and view-points whence the whole panorama could be seen, and elaborately constructed banqueting halls open to the air and covered walks,' and that when Pompey reproached him with having arranged his villa well for the summer but having made it uninhabitable for the winter, he laughed and said : 'Do you suppose I have less sense than the cranes and the storks, so that I do not change my dwelling with the seasons ?' Other authors (Varro, R.R. 1. 2. 10 and 13. 7: iii. 4. 3 : Columella, R.R. i. 4. 6; Plin. N.H. xviii. 6 § 32) speak mainly of its extent, Pliny remarking that the censors found that there was less to plough than to sweep. The piscinae Luculli which Varro mentions are not apparently to be identified with .any of the large cisterns in the neighbourhood of Frascati, but were actual fish-ponds at the villa at Baiae rather than at the villa at Tusculum. Columella, copying Pliny, simply says that Lucullus' villa was too large for the ground in which it stood, while that of Q. Scaevola was too small. Varro, however, tells us of an aviary placed under the same roof as a triclinium, so that he could see some of the birds on the table and others flying about the windows. At c. 158 of the protocollo of the notary Campana in the Archivio di Stato is a document concerning the sale in 1571 by two brothers Caro of Civitanuova to Donna Beatrice Arias de' Cinciis, wife of Dott. Evangelista Recchia, of a villa in the territory of Frascati called Villa Piscina. This is no doubt a different property, but perhaps also once belonged to Annibale Caro. See Lanciani, Storia degli Scavi, ii. 85, 86 ; iii. 50 sqq. The Villa Montalto, S. of the Villa Conti and E. of the Villa Muti, was built at the end of the 16th century by Cardinal Ottavio Acquaviva the elder, but after being for a short while in the possession of Cardinal Scipione Borghese, it passed to the Peretti family : it was bought at the end of the seventeenth century by the Odescalchi, dukes of Bracciano, so that it is sometimes called Villa Bracciano (Lanciani, Storia degli Scavi, iii. 54). In 1835 it was sold to the Propaganda Fide, and is now the property of Duke Grazioli.1 Nibby Andlisi, iii. 353, is probably right in considering that it occupies the site of an ancient villa, though hardly any actual traces are visible—only ancient concrete at two points on the N.W. side.

1 Grossi-Gondi (Tempio di Castore, 17—cf. infra, 355) publishes a Greek metrical sepulchral inscription preserved there found in thetenutaofLunghezza or that of Tor de' Sordi (Papers, i. 146 ; iii. 116). CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 251

To the S. of the Villa Montalto is the Villa Cavalletti. An important prehistoric necropolis was found in the grounds to the E. of the villa itself, where T is marked on the map. The objects are now in the Museo Preistorico at Rome. The tombs of the Vigna Giusti opposite Casale del Fico1 (really on the N.E. side of the Casale Guidi: compare our Map II. with Not. Scavi, 1902, 135, Fig. I, and see ibid, 1877, 327) on the opposite side of the Via Latina, and the tomb of the contrada Boschetto near the so-called Capanne di Grottaferrata, opposite the Mola dei Monaci {Not. Scavi, 1900, 405) probably belong to the same cemetery, and the inhabited centre to which it belonged may perhaps be sought on the summit of the hill on which now stands the modern Villa Cavalletti. A full account is given by Colini and Mengarelli in Not. Scavi, 1902, 135 sqq.: cf. Pinza in Mon. Lincei, xv. 350 sqq. The summit was later occupied by a Roman villa, scanty remains of the substruction wall of which exist : they are in opus reticulatum of selce, facing 300 S. of W. (Mattei, Tuscolo, 39 fin. 40, mentions walls ' which came close together like those of a temple ' found recently (before 1711) in a vineyard near). For late tombs found near the road see Wells, op. cit. 155. Grossi-Gondi believes (p. 101) that this may have been the villa of Gabinius : Cicero describes it as constructed at the expense of the public treasury, as being of great size, and as not being very far from his own : ad caelum exstruxit villain, De domo, 47 § 124: ad hunc Tusculani [in inonle\ montem, In Pisonem, 21 § 47 : bona ad vicinum consilient de Palatio, de Tuscula7io ad item vicinum alterum consulem deferebantur. The last is the statement in the description of what occurred after Cicero had been exiled in his Oratio post reditum in Senatu habita, 7 § 18, cf. the similar passage, De domo, 24 § 62 : eram etiatn tuo iudicio civis incolumis, cum dornus in Palatio, villa in Tusculano altera ad alterum consulem transfere- baturf senatus consules vocabant, columnae marmoreae ex aedibus meis in- spectante populo Romano ad socrum consulis portabantur, in fundum autem vicini consulis non instrumentum aut ornamenta villae, sed etiam arbores transferebantur, cum ipsa villa non praedae cupiditale {quid enim erat praedae ?) sed odio et crudelitate funditus everteretur. But, as we have seen, we do not know the site of the villa of

1 Stevenson (cod. cit. 24, 241') refers more than once to the discovery, on the N. side of the road close to II Fico, of archaic pottery by Pasquale Antini, a native of Frascati (now dead) to whom he owed much information. This was found in a natural (?) cavity in a pozzolana quarry. 252 THE BRITISH SCHOOL AT ROME.

Cicero, and the evidence for placing the villa of Gabinius here is quite insufficient. Nor is there any real evidence for placing here the hill of Corne, mentioned by Pliny, N.H. xvi. 242 : est in suburbano Tusculani agri colle, qui Come appellatur, lucus antiqua religione Dianae sacratus a Latio, velut arte tonsili coma fagei nemoris. In hoc arborem eximiam aetate nostra aniabat Passienus Crispus {Prosopographia, iii p. 14, no. 109) osadari conplectique earn solitus, non inodo cubare sub ea vinumque Mi adfundere. Vicina luco est ilex, et ipsa nobilis xxxiv peduni ambitu caudicis decent arbores emittens singulas magnitudinis visendae silvamque sola faciens. Lanciani {Bull. Com. 1884, 198) cites C.I.L. xiv. 2628, a pedestal found in the theatre of Tusculum, with the following inscription d.... \de s]enatus sententia ex muneribus Fabi C.f. Passieni Saturnini auguris aed{i) lustr{andae). But there is no adequate ground for placing the Villa of the Passieni here, or as Lanciani does, at the Villa Montalto: indeed, as we shall see below (p. 302) there are good reasons for placing it.at Frascati. There seems to be no real justification for identifying the hill of Corne with Algidus, as is done by Morpurgo {Mon. Lincei, xiii (1903) 345): see infra, 415. To the E. of the Villa Cavalletti is the modern road from Frascati to the Ponte degli Squarciarelli, which probably follows the line of an ancient road, though, as Grossi-Gondi says (p. 146), after its modernization in the middle of the nineteenth century, it is impossible now to trace its original course. Mattei {Tuscolo, 18) speaks of a piece of the pavement of an ancient road remaining above the garden of the Villa Odescalchi {supra, 250), called generally Le Pietre Liscie, on the road to Marino. Capmartin de Chaupy, however {Maison d' Horace, ii. 243), saw paving stones on the road from Frascati to Marino in situ, and remarks (in 1767) that they had been since removed; and in 1892 {Cronachetta Armellini, 1892, 178) traces of an ancient road were found near the great cascade of the Villa Conti, or Torlonia, though no details as to its direction etc. are given, so that we cannot be sure to what road they belonged ; and Stevenson, in his account of an ancient building in this villa {supra, 247) does not mention the road at all. Some way to the S. of the intersection of the modern road with the Via Latina at the Pedica, where Mattei saw the pavement of the ancient road {loc. cit.) and on the W. side of the former, Grossi-Gondi (p. 147) CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA,—III. 253 noticed in the Vigna Gentilini, a piece of ancient paving, which I also saw, which proves either that the modern road is to the east of the ancient line, or that (as I have conjecturally shown in my map) there was a branch road diverging S.W. from it and joining a short cut to the road from the Via Latina to the Aqua Julia. Rocchi, op. cit., shows it in his map, and states that traces of pavement were visible until 1897 in the lane prolonging it to the S.W. towards the Mola. Nibby {Analisi, iii. 597) refers to the section of this road from the Ponte degli Squarciarelli to Marino {infra, 280). Before we follow the Via Latina beyond La Pedica, we must return to the portion between the twelfth and thirteenth milestones. The ancient road ran, probably, straighter than the modern : some of its paving-stones are visible in situ, just before the houses marked II Fico, in the bank on the E. of the road, and also in the fieldwalls ; and a portion of its pavement was brought to light just inside the enclosure wall of the Villa Cavalletti, where some brick debris may indicate a tomb. Wells, Alban Hills, I. Frascati (p. 155) alludes to the discovery in vineyards belonging to the Cavalletti estate 'nearer the Marino road (than the villa itself) and on a line with that of Rocca Priora' of pavement of the Via Latina .(?) with traces of a branch road leading to the villa (?) and of late burials under tiles, one of which bore the stamp C.I.L. xv. 1445. b. 3 (first century A.D.). Cf. Cozza 77 Tuscolano di M.T. Cicerone in Giornale Arcadico, cxc. 115 (and separately, p. 23), who speaks of having seen the ancient road under- ground close to the gate opening on to the modern road. On its north side, just W. of the debouchement of the lane from Grottaferrata to II Fico, I saw in March 1904, in a quarry, some late tombs 'alia cappuccina' : the tiles bore no stamps. On the south of the road is a district known as Bagnara (the name is not marked on our map, but is inserted in that of Grossi-Gondi) in which stand the Casali Giusti and Guidi. To this locality De Rossi {Ann. Inst. X873, 193) refers the discovery of various important antiquities in the seventeenth and eighteenth centuries. He quotes the account of Volpi, Vetus Latium, viii. 236) who, writing in 1740, mentions the discovery of a fine statue near the villa of Cicero (which he supposed to be at Grotta- ferrata) close to the locality called Bagnara, by Cardinal Carlo Barberini,1 Cardinal from 1652 till his death in 1704, and commendatory abbot of 1 In De Rossi's quotation qtiodam is a misprint for quondam. 254 THE BRITISH SCHOOL AT ROME.

Grottaferrata from 1679 onwards ; while in 1730, Francesco Bianchi, the then owner of the ground, discovered ' innumerable fragments of worked marbles, very many tiles of the largest size, under which were very numerous dead men's bones [the superlatives are in the original]. Hypocausts and stairs of peperino were also found, leading to chambers paved with mosaic and tesselated work : the walls still showed ancient paintings in places, and were finely built of square bricks. Within the building Bianchi discovered as many as twelve marble statues, which he gave to Cardinal Melchior Polignac, a Frenchman, who soon sent them to France. Among these was found a white stone with large letters (C.I.L. xiv. 2514), and the base of a statue with the inscription following (ibid. 2517) soon appeared. All these were situated in a quite low part of the site: so that the excavators descended 20 palms (4'45 metres) below the level of the field, which was itself deep, attracted by the discovery of a very fine marble arm and hoping that they would find the rest of the statue.' Giorgi (Scked. Casanat. xvi.) gives the date of the discovery as 1731, and Ramagini (apud Muratori 353. 1. 2) states that eleven statues wearing the toga and other marbles were found. (For Polignac cf. Papers, iv, 115.) As a matter of fact, there were far more inscriptions than the two mentioned by Volpi—indeed the whole group C.I.L. xiv. 2505-2518 seems to belong to this site, as they are all given together (except 2507) by an anonymous hand in the notes of Suarez (Vat. Lat. 9140 ff. 117, 198—the following note being added on the latter leaf, ' Iscrittioni di Grottaferrata, raccommandato al Don Atanasio Gradenigo monaco di S. Basilio ivi professo—a di' 25 Maggio 1671.') Ibid. 2493 (a dedication to Aesculapius) was also found here. From the existence of C.I.L. xiv. 2507, 2509 in the Palazzo Sciarra, De Rossi infers that Cardinal Carlo Barberini had already found both the inscriptions and the statues, but had only transferred some of them to his palace, leaving the rest on the site; and he adds a note from a MS. record found by Canon Santovetti that in 1678 Cardinal Barberini transferred to his palace ' two very fine statues of white marble, one of which is a Muse 12 palms (278 metres) high, holding in the right hand a plectrum and in the left the lyre: the eyelids are of bronze and the eyes of precious stones. The face and hair are very fine, with a cloak down to the feet. (Doc. Ined. iv. 19 sqg.—inventory of 1738, p. 38, 'a statue 12 palms high representing a Muse with the lyre, in part restored . . . valued at 380 scudi.' It was CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 255 sold in 1815 and is now at Munich, Glyptothek, no. 211,1—really Apollo Citharoedus.) The second is a Faustina, 10J palms (2-34 metres) high, the right hand is beckoning, the left holds the spear: the face and marble are of great beauty {ibid. p. 25 (?) 'Julia Augusta in atto di commandare alta p. 10'). Not far off were found other statues, of which two are of inestimable value. One is a thirsty slave biting one arm, in the hand of which is a bone . . . which is held in great estimation by sculptors {ibid. p. 42, 'A seated statue 7 palms (1^56 metre) long, representing a slave biting the arm of a man, on a wooden pedestal . . . The whole valued at 101 scudi'). The second is a youth carrying a hind in his arms, 5 palms and a half (r22 metre) high, slightly bent by the effort which he makes to carry the hind in his arms {ibid, 'another statue 7 palms (1*56 metre) high, repre- senting a youth with a laurel wreath, dressed as a shepherd, with his knee on a rock supporting a dead kid, on a base of white marble, resting on a sepulchral urn . . . the whole valued at 72 scudi'). These four statues are in the Palazzo Barberini at the Quattro Fontane in the gallery of Cardinal Carlo Barberini.2' The inventory notes (p. 56, nos. 47, 63) two headless female figures with a cornucopia, and a male torso (no. 53), as found at Grottaferrata ; but they may belong to the excavations mentioned supra, 233. We must add that C.I.L. xiv. 2523 (the inscription of M. Pompeius Asper, who, however, never rose beyond the rank of praefectus castrorum of the twentieth legion) was also seen at the Abbey of Grottaferrata by Fra Giocondo, before it was brought to Rome. But De Rossi can hardly be right in supposing actual kinship between him and the Iulii Aspri; unless we assume adoption, involving of course a change of gentilicium. It is certainly true that the dates suit well, the title of praefectus castroruin having been in vogue only before Septimius Severus, being replaced after- wards by the form praefectus legionis (see von Domazewski in Bonner Jahr- bilcher, 117 (1908), pp. 119, 120) while C. Iulius Asper, who is mentioned in two or three of these inscriptions, was consul for the first time about

1 Cf. Jordan-Hulsen, Topographie, i. 3. 69, n. 75. 2 Cassiano del Pozzo notes in his diary preserved at Naples (Cod. V.E. 10) published by Schreiber, Unedirte Romische Fundberichte (reprinted from Sachsische Berichte, 1885) p. 32, no. 36) as existing at the Palazzo Barberini a Capo le Case statues of Hermaphrodites (cf. the inventory cited p. 56, no. 48) found at Grottaferrata in the Vigna Marusti. Whether this refers to these excavations [ do not know. 256 THE BRITISH SCHOOL AT ROME. the reign of Commodus {Prosopographia, ii. p. 168, no. 115) and for the second time in A.D. 212 with his son C. Iulius Camilius Galerius Asper C . . . . cius, cf. ibid. p. 184, no. 157). Grossi-Gondi, whether on his own initiative or from a misunderstanding of De Rossi, wishes to invert the chronological order, and make the Pompeii Aspri the successors of the Iulii Aspri1 (p. 99). This is, I should say, certainly wrong. A more serious question is as to the locality of the discoveries: for, as will be seen from our map, there is another Quarto Bagnara to the W. of the abbey of Grottaferrata {supra, 227), and a third in the Quarto Campovecchio {infra, 270). The remains now visible are scanty, but there is a good deal of debris in the vineyards, and walls, mosaics, vaulted substructions, etc, have, as I was told at the Casale Guidi (where various architectural fragments are visible), been found in all directions. The low ground to the S.W. of the Casali Giusti and Guidi is know still as La Bagnara ; I was told that walls and a mosaic pavement had been found there some twenty years ago, and also a marble threshold close to the path, and that the place was known as Bagni di Cicerone : the discoveries seem to have been made on both sides of the path, which is not, therefore, of ancient origin. Further to the E. rises the Colle delle Ginestre, the site selected, as we have seen, by Albert for the Villa of Cicero. Lanciani {Bull. Com. 1884, 192) is undoubtedly right in saying that this is not the site of the villa of the Aspri. The remains visible on the hill seemed to me to fall into three groups. On the north-west slope of the hill is a substruction wall of opus quadratum facing 30° W. of N., three courses of which still exist, built upon a concrete foundation, and tailing into concrete behind. The blocks are O'5o and o-54 metre high, and the faces were apparently bossed. On the W. edge of the hill is concrete facing 200 S. of W., with vaulted substructions much covered by earth. In the vineyard just below I found a fragment of the brickstamp C.I.L. xv. 911 b (first century A.D.). In the vineyard above Albert saw a cuniculus with shafts lined with cement which he took to be an aqueduct (but it was probably a reservoir) and to the E. of it a chamber with paintings : both of these are now covered up. Apparently Stevenson (cod. cit. 24V) also saw the cistern—the plan 1 De Rossi (Bull. Crist. 1872, 121) derives the name of the Fundus Ponpegi iuxta, tenimentum Cryptae ferrate of the Bull of Honorius IU. from the Pompeii Aspri, Tomassetti ( Via Latina, 151) from Pompeius Falco {Papers, iv. 119). Pompey the Great had a villa at Tusculum (Cic. Phil. xi. t;. § 11), but we do not know where. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 257 on f. 25 may perhaps refer to it. To the S.E. are remains of the main building, just at the S. edge of the highest part of the hill, consisting of walling of opus reticulatum, facing 15° W. of S., cut by the modern path over the top. Further back, under the olives near the top of the hill, I found a brickstamp, which I believe to be unpublished. CZ3 TREMMIE uo The lettering is good, and apparently of the first century A.D. On the E. slope of the hill is a platform of selce concrete facing S., marked as 1 baths' by Albert. There are buttresses 1 metre wide at intervals of 3-25 metres, projecting V25 metre from the wall, eight of which are visible : the facing (preserved only in one place) was of small opus quadratum of peperino, the blocks being about O"2i metre high and thick. The platform is preserved to the height of about six feet, and it can hardly be described as 'one of the most grandiose ruins of Latium.' (Tomassetti p. 140 note.) The terrace above it has traces of mosaic pavements. Albert discovered there a marble disk, with a double mask on one side and a marine chimaera on the other, also three fragmentary inscriptions, one possibly a dedication to Jupiter depulsor, the other two sepulchral in character (C.I.L. xiv. 2562, 2563) and the brickstamp C.I.L. xv. 1323 a (this example of it is not noted in C.I.L.). He propounded the theory that this was the villa of Cicero {Rev. Arch, xxxviii. (1879) 21 sqq., cf. PL XV. for an illustration of the disk).1 The view from the hill is a fine one, but not out of the common in this part of the country. Opposite the Villa Cavalletti a lane diverges from the road from Frascati to the Ponte degli Squarciarelli, which has traces of ancient pavement: I think a branch of it rejoined the Via Latina just to the S. of the Casale Reali, whereas Grossi-Gondi does not admit this. In any case the more important section of it ascended to Tusculum, and with this we must now deal, first glancing, however, at the villa, of which unimportant remains are now visible in the triangle to the E. of La Pedica. Grossi- Gondi (p. 103) places here the villa of Asinius Pollio, but it is clear from the map given by De Rossi {Ann. Inst. 1873, tav. d'agg. R.S.) that it was situated a good way further E. 1 He also describes (p. 27) a sarcophagus of white marble of a young girl (whose body was found within it) discovered in the Vigna Campoli, on the S. slope of the hill of Tusculum : this I cannot locate more closely. It cannot, of course, be the same as that mentioned infra, 317. S 258 THE BRITISH SCHOOL AT ROME.

Grossi-Gondi tells us (Bull. Com. 1902, 326) that in a vineyard 300 metres from the last casino of the Villa Aldobrandini due S. of point 550, a metre below some paving stones of this road, there was found the tomb of a child, covered by five pairs of tiles, two of which respectively bore the stamps C.I.L.xv. 1027 and 1036, belonging to about 123 A.D., indicating clearly that its pavement at this point dated at earliest from that period [if not, as one would suppose from the character of this burial, even later]. The discovery is a curious one, and though Grossi-Gondi believes that the paving stones (even though he admits that they had been turned over by a previous owner of the vineyard, so that they were not actually in situ) lay sufficiently near their ancient position for us to be certain that the tomb was under them originally, one must, I think, suppose rather that it lay to one side ; for it is not easy to believe that this road did not exist previously ; while that its course was changed for so trifling an object is inconceivable. The description of the locality in his Tuscolano, p. 103, n. 3, is inaccurate: 'at the highest part of the deverticulum, where it has almost reached the level,' would naturally be taken to allude to the neighbourhood of the amphitheatre ; but an examination of the passage in Bull. Com. makes the point clear. We may also note that there are now, at any rate, no vineyards E. of point 550 on this side. We may mention here the inscription of Ulpia Ephyre (C.I.L. ix. 3279) probably erected over her cenotaph at Corfinium, where it was seen, built into the cathedral of S. Pelino, by Accursius in the first half of the sixteenth century; while she was actually buried here or in this neighbourhood (condita tu pleno secura agis otia somno frigore qua gelido Tusculus alget ager, quaque via silicem terit orbita versa rotarum et dat post decimum tertia signa lapis). The exact course of this deverticulum 1 cannot at first be determined (though the remains of a tomb (?) in concrete above the letters Tu of Tus- colana make it probable that the road ran close by); but from the point where it turns to run E.N.E. the pavement (3-12 metres in width) is well preserved, and has been cleared in recent excavations, which took place in 1849-54. Here we find on its left (N.N.W.) a barrel-vaulted chamber in

1 Lanciani Bull. Com. 1884, 195, wrongly places the discovery of the inscription of Rubellia Bassa (infra, 333) near the southernmost casino but one of the Villa Aldobrandini, W. of point 550. He notes that near this casino on the N. are the outer walls of a large cistern. CLASSICAL TOPOGRAPHY QF THE ROMAN CAMPAGNA.—III. 259 concrete faced with opus reticulatum, which may have been, it seems to me, originally a reservoir, but was subsequently a columbarium with niches, each for two urns. Some fifty metres further on,are chambers in opus reti- culatum (25 on the plan, PI. XXVIII.), in the middle of which is the solid concrete of a square tomb : in an angle of one of the chambers there is a tufa cippus with a hollowed field, but without inscription. A little further on is the core of a large circular tomb (26) a mass of concrete io-so metres in diameter: the inscription belonging to it was found in 1849, and runs as follows: M. Coelius M.f. Vinicianuspr(aetor) Opsiliaitxor fecit (C.I.L. xiv. 2602). Canina {Tuscolo, tav. xxvi.) gives a view of this tomb; in Edifizi, vi. tav. lxxxiii. he gives another view, and adds the plan and section of a tomb close by discovered in 1854, with a chamber in the form of a Greek cross. From the lettering, the inscription dates from the period of Augustus at latest and we know that Coelius was Tribune of the people in 53 B.C. (Caelius apud Cic. ad /am. viii. 4. 3, cf. Bellum Alexandrinum 77) and was placed in command of Pontus by Caesar with two legions six years 'later. For the discoveries of 1875, see Wells, op. cit. 193, and infra, 334. Below the tomb is a reservoir (27) noticed also by Lanciani {Bull. Coin. 1884, 193), a single barrel-vaulted chamber, 16-50 by 600 metres inside, and very well preserved. Further down the slope again, about 200 metres from the modern road, are the scanty remains of a villa (28), a long line of substructions, once supported by a row of buttresses: at the E. end there is opus quadratum for a length of 12 metres ; and to the E. again (29) a nymphaeum with an apse at the E. end, much sunk in the ground.' This I think Lanciani is right in identifying with the site of the discovery (as indicated by De Rossi on his map) in 1849 of the fragment of a large epistyle, with an inscription mentioning some members of the gens Asinia {C.I.L. xiv. 2599) Asinia Pollionis f{ilid) Asini Celeris; Asinius Pollio films'), though the fragment was found in a vineyard, which raises a difficulty (supra, 258). The bas-relief no. 268 in the Naples museum was not found here, for it was already published by Winckelmann, long before the discovery of this inscription, which was the first piece of evidence for fixing the site of the villa of the Asinii. The inscription too seems to be of a sepulchral nature and thus gives us no right to assign this villa to the Asinii (infra, 276). Further up the hill, to the N., is a fallen piece of vaulted concrete (30). Beyond it the road ascends in a curve, the pavement being well preserved, S 2 260 THE BRITISH SCHOOL AT ROME.

(see the plan of Tusctilum, PI. XXVIII.: 1 have unfortunately omitted to mark the whole extent of it red in my map) and soon reaches the amphitheatre (infra, 338). At one point in the ascent the pavement widens out for about two metres on the upper side, the original edging having been removed.

XV.—THE ROAD TO CASTRIMOENIUM (MARINO). A little way beyond the Villa of Voconius Pollio (Papers iv. 149) the highroad to Marino is crossed, almost at right angles, by the Via Cavona (id. 125 sqq.), which from this point runs N. to the Casale Ciampino and S. to the Via Appia at Le Frattocchie. It will be well first to follow it northwards as far as the Valle Marciana (supra, 223). On the E. side of it, a little W. of the Casa Trinca, are the remains of a large villa in opus reticulatum. From this came the brickstamp C.I.L. xv. 754 b (Antoninus Pius, Faustina, or Marcus Aurelius) which we found in the vineyard below. Here there is also a small rectangular building in opus reticulatum. We saw here, too, on Nov. 23rd, 1899, the following unpublished sepulchral inscription, on a cippus of peperino CAPTIA-M L-HETAERA CAPTIA-D-L AVGE-POSILLA. It is possible that the statue with the name of Drusilla, as Venus, seated, from the Vigna Vitali, which is spoken of by Stevenson ( Vat. Lat. 10572, 19, 157v) as found at Sassone, was discovered here, as it is not mentioned in any of the descriptions of the excavations at the Villa of Voconius Polio. Further E. are some ruins in opus reticulatum cut by the railway (in the map they are marked only on its north slope). One might be inclined to suppose that the inscription C.I.L. xiv. 2480, recorded by an unknown hand in Marini's papers (Vat. Lat. 9127, 115), found near Marino, which is a tombstone of other members of the gens Captia, was also discovered near here: the same authority adds that at the same place a broken pedestal of peperino was found, on which were inscribed the words TEMPLVM SPEI. Dessau is (not unnaturally) disinclined to accept this as a genuine inscription without further confirmation. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 261

The name Quarto Sassone is derived by Tomassetti (p, 93) either from Saxa, a cognomen of the gens Voconia (following Lanciani, Bull, Com. 1884, 171), or from the heaps of ruins existing there. I should feel inclined to derive it from the peperino quarries through which the railway passes E. of point 148. In any case the name can be traced back as far as a bull of 1212. Below the Villa Maruffi some paving-stones of the Via Cavona may be seen in situ. This road is soon crossed by the ancient road mentioned in Papers, iv. 117, 1. 25 as leaving the Via Latina at the eighth mile. At point 165 it bifurcates, one branch going S.E. to the Via Castrimoeniensis at point 225, and the other ascending E.S.E. to the Colle Cimino, where it terminates, so far as I know, near the Villa degli Scozzesi {infra, 268). The first branch, which has paving-stones in the field walls, passes the remains of two villas. The first of them, to the S. of the modern house at point 200, is mainly covered by earth, and only one chamber, possibly a nymphaeum, with a shelf rather over a foot wide round the sides, is preserved. The second, a little N.W. of the Vigna Cervia, is larger : to the N.W. of it is a detached piece of construction, of five courses of opus quadratum of peperino, with concrete above, which may be either part of the substructions of a small house, or a tomb. Near the house at point 230 is a reservoir, originally with three aisles, measuring 38 20 by 8'So metres over all. Returning to the Via Cavona, we find on the E. of it the Villa B°rnabei, and to the E. of that, on the slope, the remains of an ancient villa, consisting of a large platform of opus reticulatiim with tufa quoins and brick also, facing 300 N. of W. and 3°° W. of S.: above is a low upper platform with a small cistern in it, measuring 2*74 by 3^50 metres, and only 1*65 metre in height. On the top of the hill at point 192 are the ruins of a small rectangular building of uncertain date. To the N. again we come to the edge of the hill, and here the Via Cavona, before it descends to the Mola Cavona, is crossed by the road mentioned in Papers, iv. 117, 1. 22, which branches off from that previously spoken of at point 138, at the N. end of the Colle Oliva, and follows the edge of the hill as far as the Sorgente Preziosa. The highroad meanwhile shows no traces of antiquity between points 198 and 225. On the S. of it, however, on a slight eminence, is an ancient tomb: externjilly its plan is square, but its sides are concave segmental 262 . '. —, .::... THE BRITISH. SCHOOL AT ROME. .. . curves: in the centre of the N.E. side is the door, about 2 metres wide, and or! the opposite side is a false door cut out of a block of peperino, with a double panel, 1-30 metre wide. In the centre of each of the other two sides is a pillar, and a window in the dome above. The interior is circular, about 6 metres in diameter, faced with opus reticulatum, with four niches, and a square pillar of large blocks of peperino in the centre, supporting the dome. From point 225 a path runs almost due N. along the E. edge of the railway, which is perhaps of ancient origin, though it has .no traces of paving; but it seems to be a necessary line of communication. After about half a mile it reaches the northern branch of the two starting from point 165, crosses it, and soon reaches the path along the southern edge of a branch of the Valle Marciana, a little way above the Sorgente Preziosa {supra, 222). On the descent to the spring, paving-stones are plentiful, though the cutting through which the path passes has grown several feet deeper by erosion since Roman times. The further course of the path is described (in the reverse direction) supra, 227. Returning to the E. edge of the railway, we find that the path ascends steeply to the S.E., passing the Vigna Onorati on the left: near the house (point 259) there is some debris. Above this we find, for the first time' since leaving Fontana Pesari, some definite traces of ancient pavement. . >' We then reach a large villa to the S.W. of the path, which Lanciani {Bull, Com. 1884, 192) considers to be the villa of the Scribonii Libones {infra, 267). Towards Marino is a projecting building with brick and opus reticulatum alternating, and behind it is a subterranean reservoir. Tomassetti (Via Latino., 93) notes the existence of a circular room with herring-bone pavement, with the vault fallen, repaired in places in the Middle Ages, He found in the remains of a portion of the supposed dividing wall {infra, 263) two brickstamps—C.I.L. xv. 1392 (first century), of which I have also found a copy here (presenting, however, a slight variety in the arrangement), and another, not apparently in C.I.L'., exfi. matia imeri (?), while Dessau copied here id. 1086, 15 (154 A.D.), 2270. 2 (first century A.D.) 2288.1 . !| The greater part of the remains here visible are, however, mediaeval,

1 Sepulchral inscriptions of no importance were copied here by Dessau and by Tomassetti ' in the Vigna Ingami ' (C./.L.xiv. 2478, 2490, 2491, 4228 a). • ' » CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA. — III. 263 and belong to a castle called Castel de' Paoli, described by Tomassetti (Via Latina, 89 sqq.), who also gives photographs of its entrance tower and of the apse of its church (dedicated to the Virgin, and resting upon an ancient vaulted substructure, while the external supporting wall of the apse is also of opus reticulatum) in his Campagna Romana i. (Rome 1910), p. 176, Fig. 68 and p. 196, Fig. 74. These are the only two portions of it of any importance that are preserved. We find a Massa Pauli mentioned in the eighth century Register of Gregory 11, but the reference is no doubt to Tor Messer Paolo, whereas the Ecclesia Sanctae Mariae quae posita est in loco qui dicitur Pauli of the Bull of Paschal II. (1 i 16) is probably the church on this site, for on the other no traces of a church are visible. On the other hand the Castrum Pauli indicated as a boundary of the territory of Marino in 1286 {Vat. Lat. 8034, 182, cited by Tomassetti, op. cii. p. 95, no. 6) is Tor Messer Paolo, which is crossed by the boundary between that territory and the Agro Romano, whereas Castel de' Paoli is in the territory of Grottaferrata. There is thus a good deal of confusion between the two sites. The hill of Castel de' Paoli was occupied first of all by a prehistoric necropolis, a tomb of which was found in the Vigna Onorati or Capri on its north slope, twenty metres from the enclosure wall of the Vigna Trinca: it was a cremation burial, consisting of a large dolium covered by a slab of cappellaccio tufa, and containing five smaller vases. {Not. Scavi, 1903, 202.) It is interesting to us to notice that the Colle Cimino, on the W. extremity of which Castel de' Paoli stands, has been taken by some to be the site of the ancient Castrimoenium. O. Richter (see Bull. Inst. 1885, 190) believed that he had identified two considerable portions of fortification walls on this ridge. The first, at Castel de' Paoli, in the Vigna Ingami, he believed to be part of an internal dividing wall, running from N. to S., with the remains of a gateway in it, and saw upon one of the blocks of peperino of which it is composed a mason's mark in the shape of an equilateral triangle, the side of which is 009 metre (no. 36 on the plate of mason's marks opp. p. 185). I have little or no doubt, however, that this wall is part of the mediaeval castle : its structure is quite unlike any- thing ancient known to me, though Tomassetti (p. 92) alsa believes it to be ancient in origin. The other piece of wall half way down the hill below the Vigna Galassini (not Salassini as he calls it) he describes as forming 264 THE BRITISH SCHOOL AT ROME. part of the primitive enceinte, ' The stones,' he says, f are hardly squared, but well jointed ; with the exception of a portion of the walls of Civita Lavinia, the remains seem to me to be the most ancient in the Campagna Romana.' In this piece of wall, composed of two or three courses of stones, two blocks occur in which is cut a mark like a broad arrow (no. 37), which so often occurs on the Servian walls. On the plate he gives three other marks as occurring at Castrimoenium (nos. 38-40). This second piece of wall I have never actually been able to find {infra, 267). Another argument1 in favour of the theory that this is the site of Castrimoenium is the existence, until a few years back (it has now disappeared), of a curved marble seat bearing in large letters, quite possibly of the time of Augustus, the inscription MOENIEN, no doubt part of the word Castrimoenienisium) (C.I.L, xiv, 2474). Marocco (Monumenti dello Stato Pontificio vii, 47) seems to have seen more of it, for he gives more letters of the word at the beginning, and gives its length as 18 palms (4 metres). Dessau, no doubt influenced by this, suspects that this was once the Vigna Settimi (Vinea Septimii prope Marinum),'\n which one Paolo Mancini copied for Doni several inscriptions which had recently been excavated—a dedication to the Genius Municipii by Q. Aurelius Faustinianus iiii vir (C./.L. xiv. 2454), another dedication Matri Deum ex iussu Aburius Genialis fecit {ib, 2457), a dedication to Augustus by the municipality (1-2 A.D.) (ib. 2459), an inscription recording a restoration to a public building with the munificent assistance of Hadrian (ib. 2460), dedications (a) to M. Aurelius, made in 145 A.D. (ib. 2461), (b) to M. Aurelius or L. Verus (ib. 2462), and to two individuals unknown, who had held municipal offices (ib. 2471, 2472). These comprise the majority of the inscriptions erected by the municipality as such, and all the dedications to members of the imperial house that we know of as having been made by the people of Castrimoenium. It is obvious, therefore, that the Vigna Settimi occupied the site of the forum of the city cf Castrimoenium of the second century A.D. (Tomassetti p. 97). Unfortun- ately, we cannot be certain at present of the position of the Vigna Settimi. (Another inscription, which must belong to Castrimoenium, mentioning the erection of a statue by the populus, or do, and decuriones of the (Castri)-

1 The document of 1605 cited by Grossi-Gondi, p. 198, n. 1, refers to the mediaeval fortifications of Castel de' Paoli, and is no argument for fixing the ancient city here. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 265 moenienses) was found in the parish church at Montecelio in 1853, but must have been brought there in mediaeval or modern times.) In his printed work, however, published postumously in 1731 by Gori from his papers long after his death (1647) Doni says that he saw no, 2454 in the Vigna Colonna near Marino, and Fabretti (Inscriptiones, TJ, 82) places it in the Vigna Bevilacqua below Marino (which belonged to the Colonna family). Doni states too that he copied z'£. 2461 there himself. In this same vineyard was found in 1632 a travertine slab, bearing a copy of a decree of A.D. 31 (ib. 2466), moved in the municipal assembly of Castrimoenium by L. Cornelius A.f Fal(erna tribu) Pupillus in favour of M. Iunius Silani l{ibertus) Monimus, proposing that there should be given to him a site (for his tomb) described as follows: locus qui est extra portam Medianam ab eo loco in quo schola fuit, longus pedes- Ixvii, ad rivom aquae Albanae et a via introsus, in quo antea columnar publicum fuit, latus pedes x (a site outside the porta Mediana, from the place in which the schola was, for a length of 67 feet up to the channel of the Aqua Albana, and inwards from the road, where formerly there was a public quarry, for a width of ten feet). The Aqua Albana here mentioned is very likely identical with the Aqua Augusta {infra, 394) though Nissen (Ital, Landes- kunde, ii. 582) wishes to identify it with what has generally hitherto been taken to be the Aqua Ferentina {infra, 279). Columnar, a word which only occurs here, is interpreted by Lewis and Short to mean a stone quarry. Silanus' tombstone {ibid. 2467) was found at Marino in 1790, but we do not, unfortunately, know the exact spot: the tombstone, however, gives the same measurements, and states that the site was granted by public decree. We also have preserved to us by various Renaissance copyists the tombstone of L. Cornelius Pupillus (ib, 2468), which Fra Giocondo saw in the theatre (amphitheatre ?) at Albano, and later writers in the church of S. Paolo. From this we find that he was a praefectus fabrum, and flatnen, quinquennalis, and patronus of the people of Castrimoenium. The dedication made by him to Jupiter and other deities given by various writers and mentioned by Tomassetti (p. 97, n. 2 : Orejli, no. 1393) is a forgery of Ligorio (C.LL. xiv. 129*). The Vigna Bevilacqua is also called Bel Poggio—the site is a little way to the N.E. of the modern village of Marino—and on the whole, it 266 THE BRITISH SCHOOL AT ROME. seems to me very probable that this was identical with the Vigna Settimi, and that the forum of Castrimoenium in imperial times was situated here. In that case we may, I think, suppose that the ancient town occupied the site to which, as so often in Italy, the mediaeval town returned. Certainly, too, the road from Rome seems to lead far more directly to Marino than to Castel de' Paoli. In the Vigna Bevilacqua was also seen the sepulchral inscription {C.I.L. xiv. 2485) of C. Herius Felix with its curiously pathetic text C. Herius Felix patronius) Heriae Helpinis. Hie est Me situs Herius felix coaequalib(us) cuncteis quei veixit annos viginti quei me reliquit leiberta(m) Helpineim) annorum natam xiii sibi unice caram quae illius ossa restituit (sic) post annos viginti et monumentum nobis aedificavi: qui legit discat esse pius. In the Villa Bevilacqua Volpi saw and published ( Vetus Latium, vii. 150. 15) the inscription, Kaibel I.G. xiv. 2092, <$>\aovla 'A

erenti [sic]

From this point a path descends steeply S. to point 287, on which there are no certain traces of antiquity, though it seems to mark a necessary line of communication: the modern path is cut diagonally by a drain at the point marked in the map. To the S. again, on the ascent to Marino, there are no definite indications either.- In the path going to the E. there are several paving stones loose (apparently none in situ). At the Villa degli Scozzesi, the summer residence of the Scots' Gollege 268 THE BRITISH SCHOOL AT ROME. in Rome, are some fragments of sculpture (including an archaistic Athena with the aegis, lifting up her drapery with the left hand—the height (without the head, which was let in separately) is 076 metre—a male Roman portrait head, a draped female statuette, etc.), and architecture (columns and capitals of peperino) and also the sepulchral inscription C.I.L. xiv. 2539. The exact spot at which these objects were discovered is not known, though they are said to have been found on the site, and tradition places an ancient villa here ; but there are no ancient remains now visible at the villa itself, though just below there are traces of mediaeval construction ; while to the N. are remains of two vaulted substructions, one of bad opus incertum, one of opus mixtum (both therefore of rather late date), and to the N.E. is a small reservoir, under the house W. of point 334. It is quite uncertain what course the ancient road can have taken after reaching the Villa degli Scozzesi. It almost certainly did not follow the line taken by the modern path southwards past point 319, for on that there are no traces of antiquity (nor is the path from Marino to the Mola of ancient origin), nor did it make the ascent to the E., to the culminating point S. of Colle Formagrotta. On this high point the only ruins are mediaeval or modern, though I saw there the upper part of a draped female statue. I have therefore preferred not to indicate its course even conjectu- rally, though, if it went on at all, it would seem most probable that it turned to the N. and led to the group of villas on the next hill northwards, with which it will now be convenient to deal. Taking these from E. to W., we first find two small groups of ruins W. of the Casa Raparelli, and then a very large villa on the Colle Formagrotta. Here were found two brickstamps (now in the Municipal Antiquarium at Rome, to which I gave them) which appear to be otherwise unknown. QJ) ATIMETI fi C^ES^R F Atimeti duorum Caesarum/(ecerunt) (?) |~Ij PRECIL Precilii) or Precil(iorum)

At point 341 are other ruins, in which we found the brickstamp C-IL. xv. 1324 (first century A.D.) and we then reach the Vigna Campini, the house belonging to which is built upon a small reservoir with a single chamber. Stevenson noted (Vat. Lat. 10568, 7V) in September 1883 the existence of this reservoir, which, he says, was fed by an ancient aqueduct from the Villa Schiboni. Here he saw four fragments of a bas-relief in CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 269 white marble in imitation of the Egyptian style, a large mask representing Oceanus, various fragments of columns of peperino, granite, and travertine, and some paving stones. In the Vigna Gambini was also found the fragmentary inscription C.I.L. xiv. 2558, which is apparently a part of a third copy of the inscription which we have as a whole in C.I.L. xiv. 2556. (Cf. Papers, iv. 127 and supra, 242.) Possibly the brickstamp C.I.L. xv. 1263 (134 A.D.) was found in the Vigna Campini (or Gambini—I think the identification of the two is probable) in or about 1863, though it is only recorded as having been shown to Descemet by Gambini in that year. Further on we reach the Quarto Campovecchio, and the Vigna Gavotti (Villa Lugari Spiga on the map) with the ruins of a very large villa. Piacentini {Comm. Graecae pronunc. (1750) 45) records that various reliefs were found here in the eighteenth century, when the villa was built1 and remains of an ancient mosaic pavement, and also the inscription C.I.L. xiv. 2500 : ' in vinea Gavotti, quum domus in ea pro rusticatione construeretur, plura anaglypha inventa affabre elaborata, et reliquae non modicae asaroti vetustissimi.' The inscription runs as follows : [L. Iunius M.] f. M. n. Silanu\s Torquatus honoratus ari\n. xviii \triumphalib. ornarn. q. pr. inte\r civis e\t peregrinos gener Ti. Claudi C\aesaris Augus\ti\ The person in question was a great-great-grandson of Augustus, a brother of the consul of 46 A.D., and was born about 24 A.D. In 41 A.D. Octavia, daughter of Claudius, was betrothed to him. He accompanied the latter to Britain in 43 A.D. and announced his successes in Rome. At the end of 48 A.D. he was falsely accused of incest with his own sister, in order that his betrothal might be annulled, and removed from the Senate; and at the beginning of the next year on the wedding day of Claudius and Agrippina, he committed suicide, no doubt under compulsion. Octavia was shortly afterwards betrothed to Agrippina's son, Domitius. For further particulars see Prosopographia Imperil Romani, ii. p. 249, no. 559. Piacentini also records the discovery of the brickstamp C.I.L. xv. 292 (123-138 A.D.). Certainly there are now no traces of ancient buildings at the villa itself, and Piacentini's statement may refer to the large ancient villa now known as Campovecchio, some 300 metres to the S.E. 1 The date of the construction of the villa is, it is true, considerably earlier (the inscription quoted by Tomassetti p. 145, set up by Paolo Gavotti to the memory of his brother belongs to 1569)1 and Piacentini may be only repeating what he had heard, and not quoting any good authority. 270 . THE BRITISH SCHOOL AT ROME.

Other inscriptions formerly preserved at the Villa Gavotti include a dedication by two Lucii Volumnii, Verus and Severus, to Minerva (an altar dedicated by the same persons to Diana Invicta, now in the Villa Casamorato of the Strozzi family at Florence, probably, though its provenance is unknown, came from the same place {C.I.L. xiv. 2495, 2495 a)) and the sepulchral inscription of one Italia, a child of a year and six months {ibid. 2543). There were also various sculptural fragments, some of which, accord- ing to Tomassetti (p. 144, note) had been brought from Rome—by the Lugari (?), in which case this fact need throw no doubt on the provenance of the inscriptions, the first of which was seen there in the early eighteenth century (cf. C.I.L. in loc). The inscriptions and sculptures have recently been transported to the museum of the abbey of Grottaferrata {supra, 268). Two other inscriptions mentioning the family of the Iunii are to be found in the territory of Tusculum, built into the wall of the Villa Aldo- brandini: one of them is published as C.I.L. xiv. 2720, while Lanriani, Bull. Com. 1884, 191, gives both. The ancient villa itself, however, did not occupy the actual site of the casino (marked as Villa Lugari Spiga on my map)—Piacentini merely says that the discoveries were made when the villa was built—as Grossi-Gondi and Lanciani indicate in their map, but lay somewhat to the S.E. Here is a very large platform in opus reticulatum, in the N.W. front of which is a nymphaeum, in plan somewhat resembling that under the Villa S. Antonio at Tivoli {Papers, iii. 161), with a nave 3^90 metres wide, separated by a row of columns from the two aisles, each 1*04 metre wide, which, like the nave, were barrel-vaulted : the nave alone had an apse (in which was a hole for the water-pipe), the aisles ending off straight. By a modern house I saw part of a statue with the lower legs bare, and an ornamented fringe to the drapery. In front of the platform there is, I am told, an open circular cistern, which was covered by the vineyard when I was there. The place bears the name Bagnara. Stevenson {loc. cit. supra, 268) noted, among many bricks that he examined, only one with a stamp—a palm and caduceus to the right, and the following letters to the left, the left hand portion being fractured : the stamp was rectangular— CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 271

Lower down to the S.W. under the path is a long narrow platform, and to the N.W. again are the remains of another building, the nature of which is uncertain. Three arches, about 075 metre in span and cvo-O deep, separated by pillars ro4 metre wide, are preserved ; and perhaps we have before us the dividing wall of two chambers of a small reservoir. Close by is a substruction wall of opus incertum, near which is a black and white mosaic pavement. Hence the hill falls away somewhat sharply westwards towards the Valle Marciana. Apparently the path which descends hence is of ancient origin, for one of Stevenson's informants, D'Ottavi, told him in August 1891 (see Vat. Lat. 10568, I3V) that he did not know the direction of the ancient road'which is close to Maldura, Bernabei, and Valle Marciana '• —the northern branch from Fontana Pesari (?)—but believed that the upper part of it followed the path from Valle Marciana to Campovecchio, where he said that there were ancient paving stones. In the contrada Campovecchio, in ground belonging to Sig. Giulio Cjuagliarini, remains of a rectangular building measuring 7 by 6-8o metres were found in 1908 {Not. Scavi, 1908, no). From point 225 the present highroad ascends in zigzags, while the previous road made a wide curve : the cutting through which it went is still followed by a path. This is very probably the ancient line, and I have indicated it as such on the map. The mediaeval road, however, seems, from the trend of its paving, to have run through the Valle Liccia, and, in that case, would have ascended steeply near the station, by the path which in our map is marked conjecturally as ancient, between the two villas by the railway. In the Vigna Soldini or dei SS. Apostoli, above the road on the N., were found in 1850 the following sepulchral inscriptions, C.I.L. xiv. 2469 (the tombstone of Flavia Aug. lib. Marcella, the site of the tomb having been given by decree of the decuriones of Castrimoenium) 2475 (cf. add. p. 492) 2484 (cf. ibid.). Prehistoric tombs have also been found in this locality (Pinza in Mon. Lincei, xv. (1905) 332). Hiilsen (Jahrbuck des Instituts, v. (1891), Anzeiger, p. 48) attempted to show that the Apollo Belvedere was found, not at Antium, as had generally been believed, but in the neighbourhood of Grottaferrata. On f. 7 of the sketch-book of the so-called Bramantino in the Ambrosian Library at Milan (ed. Mongeri, Milan, 1875, tav. vi.) is a plan of a hexagonal building, with the statement that ' this foundation and plan were 272 THE BRITISH SCHOOL AT ROME. on the old road which led to Marino above the property of the most reverend cardinal of S. Pietro in Vincula, and were ruined in order to remove the stones called peperino: it was there that the figure was found which was standing erect, as it is shown in the drawing. The figure represented Apollo with the bow, and was 18 palms (4 metres) high and the pedestal which you see was 9J palms high, and had carved upon it four figures in bas-relief of natural size.' On what was originally the next leaf is a sketch of the figure, and from the scale it is clear that the measure of 18 palms includes the base: so that 8J or 9 palms (about 2 metres) remain for the height of the statue ; and the sketch resembles the Apollo Belvedere. As to the exact site, ' the property of the most reverend cardinal of S. Pietro in Vincula' must be the domain of the abbey of Grottaferrata, held by Giuliano della Rovere in 1476-92 and again in 1495-1503. But Stevenson, to whom Hiilsen appealed, seemed to have some doubt as to the meaning of the term ' strada vecchia di Marino,' interpreting it as either the branch road from Sassone to Torre de' Paoli {i.e. Tor Ser Paolo) or as the branch road from the Via Latina (at the Squarciarelli ?) to Marino. Of the villas of the Scribonii and Valerii, of which he speaks as being nearest to the point in question, the former is to be sought near the Vigna Galassini [supra, 267) the latter at Marcan- dreola {Papers, iv. 152 fin.). Neither of these explanations is, I feel sure, correct, and the reference must be to the road with which we have been dealing. Michaelis, however, in the text of the facsimile of the Codex Escurial- ensis published by Egger (Sonderschriften des Oesterr. Arch. Instituts, Band iv. Text, p. 130) rejects Hiilsen's supposition, (1) because the measurement of 18 palms, he thinks, refers to the statue alone, (2) because the drawing, which he reproduces, is not really like the Apollo Belvedere. Amelung, Die Sculpturen des Vatikanischen Museums, ii. p. 265, agrees with Michaelis that the drawing represents another Apollo, which has disappeared. For the provenance of the Apollo Belvedere we thus have only the authority of Pirro Ligorio and Mercati, the latter being probably dependent on the former, who says in speaking of Antium {Taur. vol. 2= Ottob. 3364) ' nelle rovine di essa citta fu trovata una statua di quel famoso Apollo che hora veggiamo in Belvedere.' Hulsen is very likely right in maintaining CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 273 that all later statements rest on Ligorio, and that he is not more trust- worthy here than often elsewhere. It may, however, be well to cite another version of his account, preserved in an isolated leaf in his hand-writing included in a volume of the Dal Pozzo-Albani drawings now in the Royal Library at Windsor (A. 17, f. iy = Inv. 10805) 'Questo proffilo e del Nicchio magiore che contiene in se tredici altri piu piccoli, e de la Basilica passata, et quelli caratteri anchora sono contrasegni de l'ordine passato. Era la detta basilica ornata di stucchi et di statue di marmo, del che mostra chiaro i nicchi che ui erano: nel cavarui dentro ui auemo uisto trouare una statua di marmo consumatissima, et con alcuni di questi ornamenti.' 'Una' is a later insertion (in the same hand) and so are the words after ' statua,' the passage originally continued (but the words have been erased) ' di Apolline molto guasta di sorte che non seruiua a nulla, •era in atto simile a quello che hoggi dl uedemo in bel uedere, il qual fu portato da Nettuno, che hoggi non e di giouene cosi bella statua molto integra. Antichamente questo 1'uoco done fu trouata si chiamaua Ansure, hoggi Antio ouer Nettuno.' Here Ligorio seems to speak of two Apollos, the Belvedere and another like it, from Antium. Below is a restored sketch of an apse with niches, lettered ' Nicchio maggiore della basilica passata ouero Tempio, hoggi si uede un poco di esso nicchio.' After a short ascent we reach the modern village of Marino, which contains no certain remains of ancient buildings, but which, as I have said, probably occupies the site of the ancient city. Of the history of the ancient Castrimoenium we know absolutely nothing. Pliny (N.H. iii. 63) mentions it as a city of the first region of Italy, and the liber coloniarum does the same (p. 233) adding that it was munitum lege Sullana and that Nero assigned its territory to tribunes and soldiers ; but this is all we hear of it from classical writers. We may, however, reasonably identify the Castrimoenienses with the Munienses, whom Pliny (with a slight inaccuracy in this case) mentions (ib. 69) among the peoples of ancient Latium who had perished without leaving any trace behind them. It appears therefore to have originated from a camp: the name Mediana, which one of its gates bore (supra, 265) is said by Dessau (ib. p. 239, n. 1) to be a word belonging to military terminology: from Lewis and Short's Latin Dictionary, it appears to be used by Vitruvius, but •otherwise to be somewhat late. From the inscriptions we learn that it was a 274 THE BRITISH SCHOOL AT ROME. municipium, with an ordo of decuriones apparently thirty in number ; its magistrates were quattuorviri and quinquennales, and its priests were fiamines. We find dedications to Augustus and to M. Aurelius among them, testifying to its comparative prosperity in the first and second centuries of the Empire. The mediaeval and modern name Marino already occurs in the donations of Constantine to the Lateran baptistery and the basilica, of S. John in Albano respectively {Lib. Pont. ed. Duchesne i. p. 175, 1. 9 Massa Murinas, territurio Appiano Albanense,praest. sol. CCC, p. 185, 1. 13 possessio Marina praest. sol. L.). For the mediaeval history, I must naturally refer to Tomassetti, op. cit. 103 sqq. Other inscriptions belonging to Marino, but of uncertain provenance, may be briefly noted. C.I.L. xiv. 2464 (cf. add.) was copied by Fra Giocondo in the church1 of S. Giovanni or in that of S. Lucia at Marino : it is the tombstone of L Caecilius L. f. Rufus, probably tribune of the plebs in 63-62 B.C. and praetor urbanus in 56 B.C. He also saw in the same two churches, ib. 2481, 2487, two unimportant sepulchral inscriptions. Dessau, however, notes (p. 237) that Biondo (Italia Illustrata, ed. Bas. 1559, p. 319) states that many marbles had been brought from Aricia to Marino for the ornamentation of its churches, and suspects that these inscriptions may have been among them : cf. also ib. 2470, a marble stele which Stevenson saw at Marino built into a house, with the sepulchral inscription of one Seberianus rei pub{licae) ver{nd), which, Dessau thinks, may also have been transported from some other place. The inscriptions of which we do not know the previous history, which have been seen at Marino are most of them unimportant sepulchral inscriptions {C.I.L. xiv. 2476, 2477—now at Ince in Lancashire—2479, 2482, 2486—the tombstone of a female slave of Lucius Aelius, the son of

1 Fra Giocondo writes 'in aedibus S. Iohannis et Luciae' : Dessau (C.I.L. in loc.) wishes to correct to 'aede ': Tomassetti, however (op. cit. 130), distinguishes the two buildings, and describes remains of each. Opposite the former he notes the existence of a small ancient marble head with a Phrygian cap, built into the wall of a house. Stevenson saw (Vat. Lat. 10572, 57") at no. 13 of the street parallel to the Corso, a cippus with the bust of the deceased, much weathered. Carlo Pancaldi, in an otherwise worthless letter (Ricerca Archeologica intorno il tempio di Leucotea Laziale . ... in Marino, Rome, 1852) in which he attempts from philological arguments to derive Castrimoenium from the same root as mensis, mentions a figured capital in the church of S. Giovanni, with a female winged figure rising out of a mass of acanthus leaves, and a genius on each side of her. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 275

Hadrian). 2492 is a broken cippus of peperino bearing the words ad Quinctianu(m), the significance of which is uncertain, unless it was a sign-post to the property of the Quinctii. At Casa Rossa (wherever that may be) was found the unimportant sepulchral inscription C.I.L. xiv. 2450, which De Rossi saw in the Colonna palace at Marino. Marocco {Monumenti dello Stato Pontificio, vii. 47) states that here (though he gives the name as Casa Rocca) was found C.I.L. xiv. 2391, a fragment of the fasti of the Sodales Augustales of Bovillae ; but this may be an error. Boccanera also carried on some excavations in the remains of a villa on the property of Prince Colonna, in the locality known as Casal Bianco. A large polychrome mosaic was found, much damaged : it was badly executed, and represented various scenes, among which a sacrifice was recognizable. There were also found various sculptures and other objects {Not. Scavi, 1885, 78). A lead weight (384 grammes), found near the Lake of Albano in 1730, and bought by Vettori 'many' years before 1749, is now in the Museo Kircheriano. It bears the inscription : dyopavo/u,ovvTo<; TTJV /3' e^dfirjvov T. AiXbov AofiiTiavov TOV dvSoKidp^ov (?) /cal Travr)yvpuip%ou ical yvfivaaidpy^ou. (Kaibel, I.G. xiv. 2417. 1 ; Helbig, Ftthrer, ii.2 1453.) Garrucci {Dissert, archeol. ii. 79) conjectures with some probability from the mention of an dvSoKidpxv^ that the weight had been originally brought from Sicily, com- paring the word dvSo/ceia (==a balance over) in I.G. xiv. 423 (from Tauro- menium). From Marino also came a tabula lusoria copied by Doni {C.I.L. xiv. 4125. 1). According to the indications of the catalogue (1880), the following works of art in the Museo Torlonia were found ' near Marino,' no. 105 (Faun). No. 124 (bust, unknown). No. 283 (Caius Caesar in a toga). Bernoulli {Rom. Ikon. ii. 2, p. 145, no. 68) mentions a bust of Antoninus Pius at Herrenhausen near Hanover, found at Marino in 1764 ; but no further details are given. Excavations were made by Fagan at Marino early in the nineteenth century and the brickstamp C.I.L. xv. 2347. I (1st century A.D.) was found : but again we know no further details. Lanciani attributes to these excavations the discovery of the lead pipes C.I.L. xv. 7785, yjii, but, as Dressel points out, incorrectly. T 2 276 THE BRITISH SCHOOL AT ROME.

The relief of Paris and Helen in the Naples Museum (no. 268) is said to have been found in the horti of Asinius Pollio ' near Marino' {sic: cf. supra, 259). In the Atti del Camerlengato, Tit. iv., fasc. 815, are some papers in regard to an excavation carried on by Prince Aspreno Colonna in 1828. They are summarized by Tomassetti (Via Latina, 98 n.), but I have myself examined them. The locality is not more exactly described than as 'in the tenuta of Marino.' The first paper we have is a note dated May 3rd, 1828, from the superintendent of the excavations, Pietro d'Ottavi (which were being carried on in virtue of a permission already obtained in the past) stating that in the past week there had been found four marble capitals, two Ionic and two Corinthian: a fluted pilaster of marble in two pieces, ten palms (2^23 metres) high, four column drums, three fluted, and one unfluted: a marble bust, with a beard, long hair, and a garland ; and various slabs of marble. On May 13th the Prince asked for permission to remove the objects to Rome, and this was granted, a letter being sent to his agent, Don Vincenzo Colonna, with a warning that the permission had expired. On May 29th the Prince gave notice of the discovery of two bas-reliefs with figures, and horses, 5J palms high, and about four wide (r26 by C89 metre): the figures were, however, all broken, and he demanded permission to bring them into Rome. On June 2nd the guard at Porta S. Giovanni reported that two marble capitals had been brought into Rome, and taken to the Prince's palace ; but there are no further papers in the fascicolo. In Bull. Inst. 1829, 38 there is a mention of the chance discovery ' between Frascati and Marino' of a ' cassettina' (a small sarcophagus) containing a fine necklace and two bracelets of gold, a ring and a large paste, which passed to the collection of the Prince of Anglona. I know no further details as to these objects. In 1838 Prince Aspreno Colonna carried on further excavations in a vineyard known as La giostra (which means a tilting-yard—it is a name applied, e.g. to the supposed site of Tellenae—Nibby, Analisi, iii. 146), but the result is not known {Atti del Camerlengato, iv. 2799, cited by Tomas- setti, 99 n.). As we have seen, the village of Marino occupies a site with precipitous declivities on all sides except the E. On this side two modern roads leave CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 277 the town—one goes S. to Castel Gandolfo, the other N.E. to Ponte degli Squarciarelli. The former must, one would think, follow an ancient line, but retains no traces of antiquity. At the chapel known as La Madonnella, at point 363, a path diverges from it to the E., along the N. bank of the lake of Albano. On its N. side is a circular tomb, a solid mass of concrete, faced originally with blocks of peperino; and just beyond are the remains of an ancient villa in opus reticulatum with stone quoins, with a reservoir in its substructions to the N., which, like the tomb, has been partly destroyed in recent years owing to the construction here of a group of modern villas. The remains marked to the E., on the S. side of the path are, indeed, no longer to be seen: though a very little way beyond, on the N. side of it, are the scanty remains of what was once a large villa, on the ridge (marked a good deal too far E. in our map). Among the debris from it Mr. St. Clair Baddeley found the brickstamp C.I.L. xv. 1510 (middle of the second century A.D.). This is the path which is taken by Gell {Topography of Rome and its Vicinity, 16 sag.) to be the prolongation of the road from the Via Appia {infra, 290) to the site of Alba Longa, which he placed on the hill of Coste Caselle. But, in the first place, though there are traces of cutting in the rock a little before reaching Pentima Corvina, their age is quite uncertain, for the rock is a soft tufa ; and, in the second, there are no traces of early walls or early pottery of any sort on the hill of Coste Caselle. Archaic cemeteries have, it is true, been found just below it, but that will not settle the question {Journal of Philology, xxvii, 45 sag.). The old discussion as to the place of the original outflow of the lake, before the construction of the emissarium (see Nibby, Analisi, i. 100) would, since the construction of the more accurate maps now available,, have to be settled in favour of La Madonnella (where the rim of the crater is decidedly lower than elsewhere): for Gell is certainly wrong in supposing the lowest point to be near the knoll marked 381, and Nibby's explanation, that some hundreds of feet of the rim have been removed by quarrying, and that the original lowest point must be sought above Albano, is equally unacceptable. There is, on the other hand, much to be said for De La Blanchere's explanation of the object of the emissarium {s.v. Emissarium, in Daremberg and Saglio, ii. 599) that it was intended, not only to prevent the lake from overflowing, but to reduce its level to such an extent that its 278 THE BRITISH SCHOOL AT ROME. waters would no longer percolate through the lower subsoil and render the country below, between the lake and the sea, damp and unhealthy. The two most important prehistoric cemeteries of the Alban Hills are connected with the lake of Albano, one being on the north and north-east (tombs to the north of Marino, at the modern cemetery of Marino,1 in the Vigna Caracci at Prato della Corte, the Vigna Delsette at Capo Croce, the Vigna Trovalusci between Palazzuolo (or Palazzola) and Rocca di Papa, and other tombs between Rocca di Papa and Monte Cavo), the other on the west (on Monte Crescenzo and at Campofattore below it, on Monte Cucco, and in the district below these two hills known as the Pascolaro, right down to the Via Appia). For the references, see Not. Scavi, 1902, 145 sqq.—it may be noted that the two tombs found in the Villa Monte- verde, at the summit of Monte Cucco, are unpublished (p. 150, n. 3)—also Pinza's summary in Mon. Lincei, xv. 327 sqq. (those near Marino are described pp. 330 sqq., those of Vigna Caracci pp. 385 sqq., those nearer Rocca di Papa on p. 393 : cf. the addenda p. 845). Many of the objects found are now in the Museo Preistorico at Rome. Following the path along the rim of the crater beyond Pentima Corvina2 we first reach, above the rifle range (Tiro a Segno), some debris, and then, at point 515, above the Pentima Batteferro, the ruins of a villa in opus reticulatum, commanding, of course, a magnificent view over the lake and the hills : little of the superstructure is preserved, but there is a small rock-cut underground cistern lined with cement, with a main passage and two branches, and two vertical shafts descending into it. The path goes on and soon falls into that mentioned infra, 279, which before long reaches Palazzuolo {Journal of Philology, xxvii. 42). Almost immediately after leaving Marino, the modern highroad to Castel Gandolfo of which we have been speaking crosses the narrow valley which runs below Marino on the S. To the E. of the bridge (Ponte del Parco) by which it does so there are no further traces of the possibly ancient track marked in the map, and this should have been marked as falling either into the road to Castel Gandolfo or into that to Ponte degli Squarciarelli. The valley itself continues for some way further, and is beautifully wooded. At the head of it there is a seventeenth century fountain, built near a

1 See Not. Scavi, 1908, 356. Archaic tombs have also been found at S. Rocco, near the cemetery of Marino (Not. Scavi, 1903, 204 : cf. Bull, di Paletnologia Italiana, 1907, 225). 2 I shall not at present deal with the lake itself, nor with the reins along its shores. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 279 spring, which has been taken to be the Aqua Ferentina, so that this would be the site of the Lucus Ferentinus, where the independent Latin cities held their meetings (Parker, Historical Photographs, 2358-2362). Nissen, however (Ita/. Landeskunde, ii. 558), objects, and is followed by Hiilsen (in Pauly-Wissowa, Realencyclopddie, vi. 2208), that the grove was obviously not situated in Roman territory, but that, according to the accounts given by Livy (ii. 38) and Dionysius of Halicarnassus (viii. 4) of Attius Tullius' meeting with the several Volscian chiefs, it must have been near the road leading to the territory of the Volsci, i.e. the later Via Appia. The first part of the objection will not hold, for this spot would have fallen within the territory of Castrimoenium ; but I must confess that the second seems to me impossible to overcome, and that I therefore agree that the spring (or stream) and grove must be sought elsewhere—possibly the former may be identified with the outflow of the lake of Nemi below Aricia. To the S. of the Ponte del Parco I observed some years ago the remains of an ancient building lying across the path; the course of it has now been changed, and the remains have been swept away by further quarrying. To the E. is the district called Prato della Corte, and here, under the ' to' of Prato in the map, the path runs through a passage, 234 metres wide, belonging to an ancient villa in opus reticidatuni; just to the S. is a reservoir 3'89 metres in width. It is recorded that the unimportant sepulchral inscription C.I.L. xiv. 2489 was found in 1871 in the locality called Prato della Corte. To the E. of the villa last mentioned, and just beyond the boundary of the commune of Marino, on the left of the path, there is various debris, but no actual walling. Both of these sites are marked in Stevenson's map, but I have found no notes of his regarding them. Further E. again, and S. of the words I Colli, he marks other remains, but I could find nothing there. The path we have been following soon falls into another which runs due S. in a deep depression, produced by much traffic and weathering in a comparatively short time, so that in places where (as here) the soil is soft, one must be careful in inferring the antiquity of a road (though in this case I think it probable for other reasons) from the mere depth of the depression in which it runs. For until, a little over a hundred years ago, Pius VI. re-opened the Via Appia through the Alban Hills and across the Pomptine Marshes, the post road to Naples still ran by Marino and to the 280 THE BRITISH SCHOOL AT ROME. east of the lake of Albano to La Faiola and Velletri, leaving Palazzuolo a little way on the right; and from the remains of comparatively modern paving further on, and from Nibby's map, it is clear that this is the line it must have taken. This path comes from the Squarciarelli—cf. Rocchi, irr Diss. Pont. Ace. Arch, cit., who believes it to be of ancient origin, and to lead to Monte Cavo (infra, 396). It is some 3'5O metres wide in places. A little to the north, in the district known as Valle S. Lorenzo, are the remains of an ancient villa in very large opus reticulatum, to which may belong the brickstamp C.I.L. xv. 145. 5 (beginning of first century A.D.) recorded as found in the Vigna S. Lorenzo in 1727. Here Stevenson (Vat. Lat. 10572, 58V) mentions the Pozzo Calverino as a spring with ancient channels, and Rocchi (p. 240), who speaks of it as Pozzo Calvino, notes- traces of the church of S. Lorenzo and gives copies of two fragmentary inscriptions. Following the straight path southwards again, there are on the east of ft two points where the map marks ' Ruderi.' At both of these we found ancient debris (though the actual constructions visible were mediaeval), and among it the brickstamps C.I.L. xv. 1051 (after 132 A.D.) and 2235a (first century A.D.), and also the unpublished stamp 1 I P-2C IC 1=1/ To the S. again Stevenson marks remains, the nature of which is uncertain. To the E. are two groups of ruins—that to the N. is a villa, scanty debris of which is visible in a modern pozzolana pit; while that to the S. is another villa in opus reticulatum with tufa quoins, to which belongs a well preserved reservoir under the modern house, with brick-faced walls : a plan of it is here given. To the S. again Stevenson marks other *•""" remains (these are near the path mentioned infra, 390 init.).. We have now reached the limits of our examination on this side, and may, having arrived at the edge of the map, return to Marino once more. The road leaving Marino in a N.E. direction follows an ancient line, which runs almost straight to the Ponte Squarciarelli. In the first portion of it Nibby saw, between the church of the Trinita, which is within the town, and S. Rocco on the right, a mass of selce concrete which in 1816 he believed to be a tomb (Secede, v. 3. cf. Viaggio antiquario, ii. 75), but which on further examination in 1823 he believed to be the remains of a reservoir {Schede, i. 58 ; cf. infra, 289, n. 1). 'On the* road from Marino to Tusculum' Henzen copied the CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 281 unimportant sepulchral inscriptions C.I.L. xiv. 2531, 2564 (on peperino cippi). Further on, at S. Rocco, the highroad diverges from the ancient line. (It may be noted in passing that there are no traces of antiquity visible along the modern path, past the Villa S. Giuseppe to the Casa dei Frati on the road to Rocca di Papa.) There is no pavement on the latter now, but it was seen by Nibby, as the passage quoted in the footnote 1 will show; and a few stones, not in situ, are still visible. On the N.W. of the road is the Casino Schiboni ; Rocchi (op. cit.) makes the ancient road of which he speaks pass through the vineyard, and states that on the W. there are remains in opus reticulatum (p. 238). Prehistoric objects have also been found in the Vigna Schiboni, in a quarry for the volcanic stone called sasso morto, a vase and an arrow head of pale selce (the latter coloured with ochre all over): bones, also coloured, are said to have been discovered at the same time. These objects are now in the Museo Preistorico at Rome (Savignoni in Not. Scavi, 1902, 114; Pinza in Mon. Lincei, xv. 30). We must now return to the Via Cavona {supra, 260) and describe the buildings on the S. of the Via Castrimoeniensis. We may take as the boundary line between the territory which will now be dealt with and that which belongs to the Via Appia the road leading E.S.E. from Le Frattoc- chie, past the tomb called Torraccio, across the Quarto Castagnola and the Quarto Spinabella, then the avenue called the Olmata, and the boundary of the Commune of Marino up to the Villa Monteverde and round as far as the Coste Caselle (supra, 277). On the E. edge of the ' Via Doganale' (i.e. Via Cavona) leading from the Marino road to the Frattocchie, about

1 Nibby, Schede, i. 58, ' dal quadrivio della Via Latina fino a Marino la strada traversa e certamente un diverticolo antico, e precisamente quello per cui secondo Frontino si andava alle sorgenti dell'acqua Giulia che furono quelle che si trovano al di sopra del ponte di Squarciarelli, le quali sono di un' acqua Hmpidissima che non perde la sua chiarezza neppure per le pioggie autunnali. La distanza, e l'esistenza delle sorgenti non lasciano luogo a dubbio. Lungo questo diverticolo e precisamente dopo il ponte di Squarciarelli trovansi avanzi dell' antico pavimento, cioe i poligoni di lava basaltina posteriormente impiegati nella strada attuale.' Nibby, Schede, ii. 128, notes that the road from Grottaferrata through the valley to Marino is ancient, there being ancient paving stones along it, but I am not clear to which path he refers, except that it apparently passed W. of the Ponte degli Squarciarelli: for he says that at one-third of the distance is the picturesque waterfall of the Mascherone, where the factories are, i.e., probably, the Cartiera (paper-works) at point 287. He may be referring to the traces of pavement mentioned by Rocchi (supra, 253). Stevenson {Vat. Lat. 10572, 57") notes the discovery at the Cartiera Magnani of a large human mask, which had recently been sold. I presume he then went on E. of Colle Formagrotta, probably by the path past C. Raparelli (siiira, 268). 282 THE BRITISH SCHOOL AT ROME.

400 yards from the railway,a subterranean chamber 3.63 m. long was recently found : both the barrel vault and the walls were covered with painted plaster, the decorations being in panels, in which were festoons, birds and isolated flowers, but only part of the vault and the upper part of the S. wall were visible, the rest of the chamber being full of earth. In this wall there was a small niche {Not. Scavi, 1907, 221). There is nothing else of note along the Via Cavona as far as Le Frattocchie, and none of the pavement is preserved along this stretch ; but it is interesting to note that it serves as a boundary of the commune of Rome for part of its course, from point 168 southwards. Eastwards of this the boundary follows another ancient road, which ascends to the ruins of a very large villa known as Tor Ser Paolo. The name Tor Messer Paolo can be traced back as far as the eighth century, when it figures as Massa Pauli in the register of Gregory 11 ; and it appears again in the tenth, when in a description of the Fundus Casana there is mentioned the tnons qui vocatur Paulelli. (See Tomassetti, Rom. Mitt. 1886, 16 ; Via Latina, 90.) It is a question whether the waterpipes bearing the names of Valerius Messala and Valerius Paulinus, found at Casale dei Francesi {Papers, iv. 150), may not have led to or from this villa. Of mediaeval remains there is but little on this site. Various discoveries were made in this villa about the middle of the seventeenth century. Kircher {Vetus Latium, Amsterdam, 1671—but the imprimatur'is dated 1668) has the following passage : ' Praeterea spectantur etiamnum in campis passim vastissima infra terram rudera, quae ambulacra, et peristylia referunt, et si recte observentur, ilia eandem prorsus formam, quam Lucullana in Agro Tusculano exhibent. Paucis praeterlapsis annis, dum rustici nescio quod solum suffoderent, ex tremulo sono concavum subtus locum subolfecerunt, quo detecto camerarum substructiones prae- validas, praeter alia non exiguam statuarum copiam (Bartoli, in Roma Antica, 1741, p. 351 =mem. 145, in Fea, Misc. i. cclxiv., says that there were 19 statues found) et interea quoque nobile et praestantissimum monumentum repererunt,' i.e. the tablet representing the apotheosis of Homer, which was in the possession of the Colonna till 1805, and is now in the British Museum {Catalogue of Sculpture, iii. 2191), which bought it in 1819, though Tomassetti (Via Latina, 89) states that it is still in the Palazzo Colonna. Kircher gives an engraving of it by Galestruzzi, for which he expresses his CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 283 acknowledgment to Marcello Severoli, who was, he says, about to write on the subject. This engraving bears date 1658, and shows the relief already restored. He does not, however, mention the fact, which we learn from Fabretti {Columna Trajana, pp. 316, 384), that two other monuments were found on the same site—the Tabula Iliaca now in the Capitoline Museum (Sala delle Colombe, 83) and the bust of Claudius now in Madrid, which was given by the Colonna family to Philip IV—the so-called apotheosis, once borne by an eagle, which is now lost (Bernoulli, Rom. Ikonographie, ii. 1, 337, no. 29). Fabretti gives an engraving of the latter, '—cuius exemplar pagina seguenti adumbramus, elegantius a praestantissimo Marcello Severolio excipi iussum et eius de Apotheosi tractationi insertum.' The dissertation of Severoli, however, for which these illustrations were prepared, never appears to have actually seen the light, for we find Gisbert Cuper stating in 1683 in his Apotheosis Homeri—cf. the reprint of it in Polenus, Utriusque Thesauri nova supplementa (Venice, 1737) ii. 28 C—that Severoli and others had never yet written on the subject, which he was therefore undertaking to do himself. Nor does either Schott, ibid. 298, who first wrote in 1714, or Polenus in his preface, allude to Severoli's work. Fabretti further states {pp. cit. 316) that the discovery of the Tabula Iliaca was casual (' non multis ab hinc annis inter rudera et parietinas lateritiorum fragmentorum iacentem, quamvis terra et arena ilIi adhaerente vix conspicuam, agnovit, et inter pretiosiora sui Musei habuit D. Arch- angelus Spagna Romanus, . . . nee diligentiae pepercit, quo parte deficiente potiretur; sed nequidquam labor plurium operarum, eiusque solicitude cessit'). According to Winckelmann indeed (Kunstgeschichte ix. 2, § 35 = Storia dell' Arte ed. Fea, vol. ii. p. 215) the Canon was out hunting when he found it. From him it passed to the Spada family, who presented it to the Capitoline Museum. It is very possible that the Colonna family excavated after this discovery had been made, and found the other objects. As to the remains in which it was discovered, Kircher supposed them to belong to a villa (containing a temple) of Claudius, while Fabretti attributed them to the sacrarium gentis Iuliae at Bovillae. That the reference is to this site is perfectly clear from the map of the Dorsum Praenestinum et Tusculanum, added to the second edition of the De Aquis (1778) of the latter opp. p. 90, in which he marks it 'BOVILLAE hod(ie) 284 THE BRITISH SCHOOL AT ROME.

T(orre) di Re Paolo' (a name which it also bears). That Bovillae was situated on the Via Appia is known to us from later discoveries—cf. C.I.L. xiv. p. 230. Bartoli (loc. cit.) states that Cardinal Francesco Barberini ' many years afterwards continued the excavations, though owing to differences between him and the Conestabile Colonna, the work lasted only a short time: still, even from that small excavation, he brought back to Rome very fine pieces in the fine Greek style. A part of the building, circular in plan, was found, at the entrance of which there were two statues, one of a woman standing, and the other of a man, broken, on the ground, in front a sarcophagus, which was left in the same place, which was under the stairs ascending to the upper floor at the sides of the theatre. A remarkable fact was, that the small lateral chambers were all lined with very thin plates of copper, fixed to the wall with nails of the same material. A very large courtyard was found, all paved with mosaic, with gutters through which the water was made to run for the adornment of the place, of blocks of peperino 10 and 12 palms (2*22 and 2"66 metres) long. A quantity of underground rooms were also discovered, which were believed to be the habitations of the slaves, where very fine large masses of marble were excavated. It is said that very beautiful things were also found in the time of Paul III. It must be noted, that in the excavations made by the Constable so many fragments of figures, such as noses, fingers, tips of toes, and drapery were found, that carts might have been loaded with them. From this the richness of the place may be conjectured.' A fine marble pavement, with a design in opus sectile (not true mosaic, but pieces of coloured marbles) representing the origin of Rome, found near Tor Messer Paolo in 1837 in an unauthorized excavation, is mentioned in Bull. Inst. 1838, 112 ; and the documents relating to it are in Atti del Camerlengato Tit. iv. fasc. 2768. It was sequestrated by the tribunal and handed over to the Colonna family in 1857. It is still in their palace, and has been described and illustrated by Tomassetti {Rom. Mitt. 1886, 3 sqq. and Taf. i.). The inscriptions found at or near Tor Messer Paolo are the following1: C.I.L. xiv. 2426 (an altar of peperino found in 1853, bearing 1 C.I.L. xiv. 2420 is given by Amati in two places in his notes as having been found in the ' tenuta del de Pauolo, vicino a Boville'; but in others he assigns it to Latera near the lake of Bolsena, with another inscription which certainly belongs to that locality ; and the latter indication is probably correct: see C.I.L. xi. 2916, 2919: Eph. Epigr. vii. 1256. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 285 the inscription Herculi Augiusto) sacrum Delfus Caes{aris) n[ostri] serbus ber(na) disp{ensatof) v.s.l.m.) ; 2431 (a sepulchral inscription found ' near the contrada torre Messer Paoli' (see Grossi-Gondi, p. 198) running thus: D.M. Claudiae Priscae coniugi piissirnae Eutyches Caes. N. ser. Tryphonianus disp{ensator) vill{ae) Mamurranae. From this inscription we learn that the villa of Mamurra, the well known favourite of Caesar, had become imperial property under the Claudian emperors; and the discovery of the bust of Claudius here may lead us to suppose that the villa of Tor Messer Paolo is the actual villa in question); 2432 (a sepulchral cippus found in the Vigna Trovalucci—not that mentioned supra, 278, but another—near Frattocchie and the Via Appia —this fact shows that it was found near Tor Messer Paolo, and not near the Castel de' Paoli to the N.W. of Marino, which the indication of Cozza might lead one to suppose: cf. C.LL. in loc. The two are indeed not infrequently confused—as for example by Tomassetti, Via Latina, 94). Marocco, Stato Pontificio, vii. 46, records the discovery in the territory of Marino, ' in contrado terre (sic) Messer Paoli' (which he expressly distinguishes from the Vigna Ingami) of the dolium stamp C.LL. xv. 2506. A report of 1840 {Atti del Camerlengato, Tit. iv. fasc. 3031) deals with the •chance discovery of paving stones belonging to the ancient road (the Via Cavona ?) and of a door-post and the half of a basin, both of white marble, by Giovanni Battista Pellini: cf. Tomassetti, 100 n. ' At Tor Messer Paolo ' (without further details) C. L. Visconti noted the discovery of the brickstamp C.LL. xv. 122, c. 13 (slightly before 123 A.D.), and I have myself found a fragment of ibid. 1239 a. (1st century A.D.). Excavations carried on by G. B. Guidi in the tenuta of Tor Messer Paolo at the end of 1852 led to the discovery of an ancient villa. A life- size male statue in marble and other fragments were found in the ruins {piss. Ace. Pont. Arch. xv. p. xxv.).1 The actual remains now visible at Tor Messer Paolo are scanty, though the villa itself must have been immense, and commanded a fine view : they consist of two platforms, the tower from which the locality takes its name being on the lower one. 1 I suspect that the statue of Domitia Longina (?) as Venus (Ny-Carlsberg 541) was found here or hereabouts. According to the Catalogue it was found ' at the Frattocchie near Albano in the villa of the Flavii '—a somewhat vague description, and, if the reference is to the villa of Domitian {which is at Castel Gandolfo), self-contradictory. 286 THE BRITISH SCHOOL AT ROME.

The construction is of various styles: we find loose blocks of opus quadratum, and several styles of opus reticulatum, besides mediaeval additions. In one place is a small reservoir in the platform, 2-37 metres square, in another a second reservoir with two chambers. In the southern portion there are large vaulted substructions, and a cryptoporticus 470 metres wide. Columns of porta santa and other marbles are still to be seen lying about. About 200 metres to the N. of point 207 and the same to the W. of point 209 is a detached reservoir (not marked in the map) with three chambers, barrel-vaulted. This seems to have supplied, not Tor Messer Paolo, but a villa on the hill just to the W. of the Via Cavona, a little to the S. of point 172 : there is much debris on the site, but nothing else. At the north-east angle of the villa at Tor Messer Paolo, near point 221, is an important junction of ancient roads. One of these comes from the Via Castrimoeniensis at point 198: its prolongation runs southwards, with one or two bends,1 to the Torraccio, a large tomb at point 192, of which Labruzzi gives a drawing (II. 33 : cf. Me"l. Ecole Francaise, xxiii. (1903), 33). It is square, with a passage i"2O metre wide going all round the central core of concrete. Just to the S. of the villa of Tor Messer Paolo, at Costa Rotonda, in the Vigna Quagliarini (cf. p. 271 for land belonging to the same proprietors elsewhere) there was found a large block of marble with an inscription as follows: salbis dd. nn. Costantino Aug. et Constantino et Costant\io~\ nobilis- simis Caesaribus felices domini fundi, an acclamation dating from after the death of Crispus (326 A.D.) and before the proclamation of Constans as Caesar (332 A.D.). I very much doubt whether Grossi-Gondi is right in supposing that felices domini fundi can be an error for the ablative, and that the domini referred to are Constantine and his sons. Remains of a large building in blocks of local stone were also visible {Not. Scavi, 1908, 357). The road we have been following eastwards from point 168 on the Via Cavona bifurcates almost at once, at point 221. One branch goes on E.S.E. Its antiquity is probable, though not certain, and it should be marked in

1 The modern path is incorrectly marked as going straight, it really corresponds with the ancient line as given in our map. There is on this no pavement in situ, but there are large paving stones in the fieldwalls. It should further be noticed that in the whole course of the succeeding description the Torraccio referred to is the tomb at point 192, not that at point 206 on the Via Appia. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 287 the map as conjectural. In any case it did not exactly follow the modern line, as we find an aqueduct C45 wide running S.E. and N.W. across the path near its beginning. On the N. side of it (not marked in the map, but just under the letters ' am ' of the word ' Campofattore') are the remains of a large villa. The path going N.W. has been in part suppressed since the map was made. The Villa Castruccio probably occupies an ancient site also, though there is nothing actually in situ to be seen—only architectural fragments, which may have also been brought there from the villa just mentioned. According to the official reports of 1905, in the Quarto Campofattore, in the Vigna Castruccio, are the remains of a Roman villa, in brickwork and opus reticulatum. Architectural fragments were to be seen, and a marble statue representing Aesculapius, i"4O m. in height, was discovered {Not. Scavi, 1895, 423): the type was that of the Naples statue (Baumeister, Denkmaler, i. p. 139, fig. 148). The reference is possibly to the villa N.W. of point 249. In May 1906, at 56 metres W.N.W. of Casale Liccia, a male statue, with the lower part of the body and the legs draped, was found under the path leading thither from Tor Ser Paolo. Further excavations, in April 1907, led to the discovery of walls in opus quadratum of peperino, and in opus reticulatum. A room with a hypocaust was found : the pilae were each formed of eleven bricks o-22 metre square, bearing the stamp C.I.L. xv. 595b (period of Hadrian). The pavement was formed of large bricks 060 metre square and o-o5 metre thick, over which was a layer of opus signinum, O'i2 metre thick, which formed the pavement. A late tomb, formed of sjabs of peperino and tiles, had been inserted into this room. A water channel, O'9O metre high and C48 wide, lined with cement and floored with opus signinum, was found, and also a smaller channel in cement O'O9 metre square, covered with bricks laid flat. A threshold, with a fine black and white mosaic in front of it, and other walls and pavements were also discovered, and the building was no doubt a villa of a certain importance, some fine painted stucco having also come to light. Fragments of columns and Doric bases and capitals now at the Casale Liccia may have belonged to this building. The other brick- stamps found {C.I.L. xv. 557, 565 d or e, and 2227) belong, except the last, which maybe earlier, to the period of Hadrian, but most of the walling is in simple opus reticulatum and perhaps earlier. {Not. Scavi, 1907, 214 sqq.) 288 ' THE BRITISH SCHOOL AT ROME.

As these remains (which have now been covered up) ran across the modern path, the ancient road must have run somewhat differently. Between the Casale Liccia and the Vigna Gabrielli there is a tomb- stone of the 'boiler' shape (see Papers, iv. 118) built into a wall, bearing an inscription of which I could make out but little : I give what I was able to read. The whole stone is 1*40 metre in length and 078 in height, and the tablet, on which is the inscription, is O'28 metre wide and 046 high. D M L NENIO CLE NI V FIL O k FECIT In the Vigna Gabrielli there still exists a peperino base with a dedica- tion to Hercules Invictus (C.I.L. xiv. 2455): another altar of the same material, seen there with the first in 1749 (when the vineyard had recently changed hands, having formerly been the property of Giovanni Battista Valentix) is no longer there. It bore the inscription Laribus, cofijegis con\- ferentibus [mag(istri) . . . (ib. 2456). In the same place is an inscription of one M. Aurelius Iulianus a rationibus et a memories, {ib. 2463—vi. 1596) erected by his son-in-law Sextus Pedius Iustus {Prosopographia iii. p. 20, no. 155) which is not sepul- chral but honorary, inasmuch as he afterwards became Praefectus praetorio •{Prosofographia i. p. 209, no. 1268). Inscriptions erected by Pedius Iustus to his father and mother are recorded (C./.L. vi. 1597, 23891), the former was in the collection of Fulvio Orsini, the latter was found in 1788 near the church of S. Carlo a' Catinari in Rome. It is not impossible that the first (if not both) of these came originally from Castrimoenium, for we know nothing of the circumstances of the discovery of either. There is also a terminal inscription of the Republican period C.I.L. i. 1126 = xiv. 2488). . In agro P. Paacili terminus totus est conlocat{us), repeated twice on two sides of a peperino cippus. To the S. is the Vigna

1 There is, indeed, over the gate of the oliveyard N.W. of the Casale Liccia the inscription Quintilianus de Valentibus Abbas pro fundi tnerito MDCCX p{osuit). CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 289

Lecce, where Ottaviano Lecce found in 1834 a statuette of a nude Bacchus, a torso of a statue, two small columns, a foot of a marble table, and a series of twelve steps of peperino {Atti del Camerlengato, iv. 2105, quoted by Tomassetti, 99 n.). W. of the Villa Capri there are some pavingstones in the path. In the Villa Capri (formerly Camporesi, and in the early nineteenth century Schrimitii or Ferrari) is the sepulchral inscription of L. Sextilius Satyr, lictor curiatius, his second wife, and his two young sons. {C.I.L. xiv. 2522.) A similar inscription {C.I.L. xiv. 2521), set up by the same man to himself and his first wife, was found in Rome in 1839, but had probably been transported thither in mediaeval or modern times. Other inscriptions, on the other hand, seen in this vineyard by Fea, belonged originally to Rome: they were not brought thence directly, but had previously been in the collection of Cardinal Passionei at Camaldoli below Tusculum [infra, 374) : see C.I.L. xiv. i8o*c. To the E. of the Villa Capri this road falls into a path, probably (though there are no certain traces) of ancient origin,1 which descends through the deep valley to the S.E. and S. of Marino, keeping on its S. side. On this side are extensive peperino quarries, close to Marino station, and perhaps of ancient origin, though still in use (PI. XXV. Fig. 2).2 On the opposite side, and cut by the railway, are the remains of two villas of no great importance—one is just W. of the station, the other a little further on above the line. The path up from the station to the town contains pavingstones, probably not in situ. To the E. of the chapel marked in the map there is a cutting which is possibly of ancient origin. Beyond the junction near Villa Capri it passes round the N.W. spur of Monte Crescenzo, on which are the remains of three buildings : at point 342 is debris, to the S.E. of it is a reservoir, while at point 374 are traces of another villa.

1 I imagine it is to this road that Nibby refers in the following passage in his notes of 1823 (Schede, i. 59) ' alia sorgente (dell' acqua Ferentina) conduceva la strada o diverticolo scoperto di recente presso Monte Crescenzi ehe diramandosi dall' Appia metteva nella Trionfale salendo per la via odierna (giudicherei piuttosto che sepolcro un' avanzo di conserva quell' opus emplecton che si trova sulla via trionfale a sinistra per chi va da Frascati a Marino fra la chiesa di S. Rocco e la Trinita'). This last church is actually within the village of Marino, and contains a picture representing the Trinity by Guido Reni. In his previous notes of July, 1816 (Schede, v. 3, 4) he had believed the mass of concrete of which he speaks to be a tomb, and spoke of the road as leaving the Via Appia near Bovillae, and passing through Marino and under Rocca di Papa to Palazzuolo ; but in the Analisi (i. 62. 114) he follow's Gell's view, and brings the road up by Monte Cucco {infra, 290). 2 See Parker, Photos 2356, 2357. U 290 THE BRITISH SCHOOL AT ROME.

Below Monte Crescenzo, to the E. of the Casa Trinca, are many pavingstones in the fieldwalls, and a little further on we come to a piece of pavement in situ, about 3 metres in width. The path is then (in the Quarto Spinabella) joined by two other ancient roads, one from Tor Ser Paolo, and one from the Torraccio. (Shortly afterwards a boundary wall marking perhaps the line of another ancient road branching from it—this would be the road mentioned by Gell, Topography of Rome, 17 : see Ashby in Journal of Philology, xxvii. 45, 47)1 ascends almost due E. through a depression, which gives it an easy slope up (cf. supra, 277) to the edge of the extinct crater which contains the lake of Albano. Along this rim there must have been an ancient road also, but no traces of it can now be seen.) Then it leaves the modern path on the W. and can be traced across the fields, passing between the ruins at point 259 (the solid concrete core of a tomb) and some peperino blocks in the modern path, belonging to the foundation of a building now destroyed. Shortly before reaching the Via Appia it rejoins the modern path, leaving the remains of an ancient villa on the E. Some way to the E. of this, at point 318, on the line of the boundary between the communes of Marino and Castel Gandolfo, are the scanty remains of a villa (indicated by Canina, Ann. Inst. 1854, on the extreme left of his plan of this district, though, if he is to be trusted, he saw very much more than is now visible) and, to the S.W. fronting on to the Via Appia, are the remains of another very large villa, in which a large crypto- porticus 2'35 metres in width can be recognized : this edifice is believed by Canina to have been the Villa of Clodius (Ann. Inst. 1854, p. 97 : cf. Journal of Philology, xxvii. 48). The discovery there in 1819 of the tombstone of M. Pompeius Crescentius with an inscription in Greek {Kaibel T.G. xiv. 1958) revealed the origin of the name Monte Crescenzo, which can be traced back to the bull of Agapetus II. of 955, confirming the possessions of the church of S. Silvestro in Capite, in which a church of S. Angelo, in monte qui vocatur Cresenzuli is mentioned.

1 In the vineyard of Sante Limiti (which is on Monte Crescenzo, cf. Pinza, Mon. Lined, xv. 335) in 1877 weie found a number (at least thirty) of late tombs ' a capanna ': one of the tiles bore the stamp C.I.L. xv. 966 b (first century A.D.). They were aligned with a road, of which the pave- ment was found, which wa-; believed to be that leading to the ancient Alba Longa. (Not. Scavi, 1877, 208) Stevenson copied the stamps ibid. 548 a. 14 (123 A.D.), 1465 a. 6 (first century A.D.) on the tiles from a late tomb found on Monte Crescenzo, possibly in this very vineyard. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 291

Near Monte Crescenzo was found a sepulchral inscription erected to two sons by their parents, the father, Apollonius, being a tabellarius Caesaris nipstri) and the mother, Octavia, a negotiatrix (C.I.L. xiv. 2465). The other branch from point 221 runs S.E., and there are pavingstones in the fieldwalls. At point 252 it is crossed by a path ascending from the Torraccio, the western portion of which is probably ancient, as there are a few pavingstones lying loose in it, one or two of which are very large. In the path I found the brickstamp C.IL. xv. 171 (about 138 A.D.) and in the fieldwalls is debris of a building. To the N.E. of point 252, however, it presents no traces of antiquity. The path going S.E. passes on the E. the remains of a mediaeval building built of small blocks of tufa (there is, however, much ancient debris about) and in the Quarto Spinabella falls into the road already mentioned from the valley below Marino to the Via Appia, leaving on the E., just before the junction, some ruins in peperino. To this point comes another path from the Torraccio, which is probably also of ancient origin, though neither E. nor W. of point 227 are there definite traces of antiquity in it. To the W. of point 227 is a small reservoir, and further on again are the foundations of a villa, in opus quadratum and opus reticulatum, soon after which we reach the Torraccio. From point 227 another path runs W.S.W. which coincides with an ancient road, some of the pavement being actually in situ in the bank at the side : it crosses the Via Appia a little N.W. of the ruins of Bovillae, and can be traced as far W. as Falcognana, on the ancient road to Satricum (Conca). A path diverges from it to the Torraccio to the N.W., in which there is pavement still in situ, while numerous pavingstones are to be seen in the fieldwalls. By it are two uninscribed cippi of peperino, each about O"4S metre square. It runs more or less parallel to the path going to the Torraccio from point 227. Just after this divergence there are remains of buildings, marked in the map on each side of our main path. Among those to the S.E. of it I saw a fragment of a block of peperino with the following inscription badly cut on it 'iVNDI IPERIM T

u 2 292 THE BRITISH SCHOOL AT ROME.

The block was 0'48 metre thick, and the part preserved was 046 metre wide and O'3S high : the letters were O"37 high.

XVI.—THE TERRITORY OF TUSCULUM BELOW FRASCATI AND FRASCATI ITSELF. In dealing with this, it will be well to describe first the course of the Via Cavona from Casale Ciampino to the Via Labicana, inasmuch as we have taken this important ancient road (see Papers, i. 176, 236, 240, 242 ; iv. 117, 125 sqq., 133) to mark the boundary of the first section of the description of the Via Latina. The portion of it to the S. of the Casale Ciampino has already been dealt with (supra, 260). The road, as has been said, divides the Vigna Senni from the Vigna Gentilini, passes over the top of the tunnel of the Frascati line (where is the Botte S. Andrea, a modern reservoir over an excellent spring of fresh water) and then descends steeply to the valley of a tributary of the Marrana Mariana. At the beginning of the descent, on the E., is the brick debris of a building overlooking the valley, and at the N.E. end of the tunnel, in the cutting, is a wall. Further N., on the hill to the E. of the road, a well shaft (marked W in the map) was found in 1907, which I saw. The slab at the top had originally served as a cornice block: it had a neatly cut round hole in it, and this was covered by a flat slab. On the W. edge of the hill above the railway, as Stevenson notes ( Vat. Lat. 10572, 40), is a rock-cut drain, and below in the cutting lies some debris, among which I found an unpublished brickstamp of the gens Naevia (first century A.D.) CZ] C • NiE • BA. (C. Nae(vi) Bal...) The object at the right hand of the stamp seems to be an amphora. It is probably a variety of C.I.L. xv. 1324 ( Naevi Bal (....)). Stevenson noted the discovery (and destruction), when the railway was made in 1890-1, of walls of opus incertum and of pilaster capitals of peperino, with a bust issuing from acanthus leaves : another capital with female head between two volutes was also found : at one extremity of it was a staff with a serpent twined round it, at the other a male bust. Other fragments included a child's sarcophagus in sperone, 1*15 metre in total length. In the upper stratum, of pozzolana, was a regular network CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 293 of cuniculi, down to 4 or 5 metres below the surface : in the lower stratum was one passage 7 metres down (still visible), but the cuniculi at the lower level did not apparently extend further. The official report speaks of it as a Roman farmhouse: a small ' Etrusco-Campanian' vase with black varnish was found. {Not. Scavi, 1891, 229.) Some 400 yards further N.E., on the hill marked 164 m. above sea- level, the railway cut through the concrete foundations of a series of ancient walls already razed to the ground, for a length of some forty yards, probably the foundations of a Roman villa, though no objects of value or architectural fragments were found. {Not. Scavi, loc. cit.) There is still debris on the ground above, and some peperino blocks running N.N.W., and a few pavingstones a little to the N. in a vineyard : Stevenson {loc. cit) notes peperino blocks and bricks, and some hundred rough paving- stones, not worn—possibly the pavement of a yard. The Via Cavona now reaches the Fonte Vermicino and crosses the Via Tuscolana, as to the antiquity of which, from Rome up to this point, there is, as I have said {Papers, iv. 51), considerable question. Beyond this point, however, I think it very probable that it represents an ancient road : it seems to be a necessary line of communication, prolonging the road from Torre Nuova, which is almost certainly ancient {Papers, iv. 146). Stevenson records, it is true, that when the railway was made, it was found to be of rough pieces of selce, and to give no indications of antiquity. The last of the six lines marked as intersecting at this point is the road following the aqueducts {Papers, iv. 118). The Aqua Claudia and Anio Novus pass underground just at this point ; but I shall deal further with them in a separate work on the aqueducts, which I hope will shortly be completed, and which will be published by the Clarendon Press. In the Vigna Guerrini, which is one of the Vigne S. Matteo, just E. of the Fonte Vermicino, Pasquale Antini, who was so often Stevenson's informant, copied the inscription C.I.L. xiv. 2613, which Dessau had failed to see in 1880. See Stevenson (Vat. Lat. 10572, I4iv). Only the last three lines are preserved, which run thus: et Hispania ulteriore provincia Baetica Municipium Concordia Iulia Nertobrigenses publice. It was first recorded by Fabretti as in the estate of one Cavaliere Silva, the owner of the old osteria of Vermicino, which was, I presume, the house at point 134 (Mattei, Tuscolo, 83 adds the spurious inscription C.I.L. xiv. 210* and another which really belongs to the district of Milan—ib. v. 5604, xiv. 224*). 294 THE BRITISH SCHOOL AT ROME.

The sepulchral inscription C.I.L. xiv. 2540 was also found in Cavaliere Silva's property, though Dessau leaves out this particular, so that the ruinae veteris oppidi (whatever they may be) are rightly localized in Papers, iv. 147. The mention of the chapel of S. Francis Xavier fixes the provenance of ib. 2737 (Mattei actually speaks of Vermicino) to the same place, and they are wrongly separated in the Corpus. The last named is an inscription in elegiac couplets in memory of Rhanos, a girl who died in childbirth before the age of sixteen. Antini also copied here the following unpublished in- scription, and took a rubbing of it, which Cozza gave to Stevenson in 1888. C ' AT) VIX ATTIA-O\ Finally, Guidi copied at Frascati in 1881 the following inscription on a lead pipe said to have been found in the Quarto S. Matteo near Vermicino Claudius Felicisshnus fecit (reversed) C.I.L. xv. 7837. c. 2 (cf. Papers, i. 200, n. 1). Wells, Alban Hills (i. Frascati, 17, 98), states that the Casale S. Matteo is built on extensive ruins, to which this pipe may have belonged. Her description (p. 100)-fixes near here (probably the actual Casale is that at point 153, to the W.) the ' podere de' Borzari,' the site of the discovery of some sculptures, consisting of a nude youthful male torso, a head of a Cupid and a head of a Satyr (this last probably belonging to a group, in which another figure grasped it by the hair, there being a hand attached to it) and a fragmentary statuette of an Amazon resembling Clarac 810 A, 2031 C, which Braun described as in the possession of Antonio Santovetti of Frascati {Bull. Inst. 1857, 67). In the Atti del Camerlengato, Tit. iv. fasc. 1036, we find various documents relating to the discovery of two sarcophagi not far off. A letter from one Celestino Benedetti, of January 28th, 1829, announces that his father, Michele Benedetti, of Frascati, in continuing the breaking up of some ground in the district called Colle Giudice, adjacent to the vineyard of the Viti family, which had previously belonged to the Aw. Borzari (Wells, op. cit.)y had found at a depth of about 4 palms (O'89 metre) a marble sarcophagus with its cover broken : it was more or less oval in shape and measured about 9 by 3 palms (i-99 by o 67 metre) and 2 (C45 metre) palms in depth with a bas relief of animals and figures. It had been conveyed to Frascati to protect it, and permission was asked to continue excavation in the hope of finding more. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 295

A further declaration before the Governor fixed the actual date of discovery as 24th January. The sarcophagus was described as having strigil markings, with a vase in the centre. At the angles were two lions, each devouring a lamb, and two human half-figures : the back was rough. It was then stated to have had no cover. The sarcophagus was well pre- served, but the work mediocre. It was decided that Benedetti might go on digging ; and on January 14th, 1830, he gave notice of having found another sarcophagus with its cover, which was broken into several pieces. It was 9! palms long, 4 deep, and 4 wide (2-li by C89 by C89 metre). At the corners were lions devouring calves, and figures with spears, and in the centre another figure. This again was not thought of sufficient importance for the Papal museums. Wells also informs us that the Faun of Praxiteles of the Lateran museum (no. 150) was found here: of this Benndorf and Schoene (p. 90) know nothing. In the Atti del Camerlengato, Tit. iv. fasc. 985, are other papers relating to excavations in this district. We learn that on February 4th, 1829, Antonio Coronaldi of Frascati was given permission to make a small excavation in the Tenuta of Vermicino, in the territory of Frascati, having obtained permission both from Count Cosimo Conti, the owner, and Camillo Poloerosi (?), guardian of the Campana heirs, the tenants. There is, how- ever, no record of what was found, though the Archaeological Commission was advised that work was to begin. Other excavations were carried on in the Campana property by Suscipi in 1833 and Pieri in 1834 and 1835 (ib. 2102, 1835): the former had heard of a treasure, and obtained leave to search for it, but we do not hear of the results ; while the latter began work with the same idea, but very soon closed the excavation, having found nothing. From ib. fasc. 3857 we learn that in February, 1854, Marchese Giovanni Pietro Campana himself obtained permission to excavate in corn- lands belonging to him at Quarto di Maria, and Colle Pisano, and vine- yards at Imagine Nova, Colle Giudice, and Pescareccia. On April 20th he reported that he had found nothing but two marble heads, some small cornices (also some large square pieces) of rosso antico and a few fragments of decoration. These were found at Grotte Piattella : cf. Papers, iv. 144 for another account of the same excavations : there, too, the district to the N.W. of the Via Cavona as far as that point is dealt with(cf. Grossi-Gondi, 176); while beyond that we fall into the territory of 296 THE BRITISH SCHOOL AT ROME. the Via Labicana {Papers, i. 240: cf. the addendum in Papers, iii. 207). I may add that a reservoir has recently been found just N. of the house marked Micara. On the S.E. of the road was a mediaeval tower of the thirteenth (?) century, known as Casamari (described by Tomassetti, Via Labicana e Prenestina, 36 note) which has recently been demolished, a modern house having been built there. I saw in 1909 a block of marble with a fragment of an inscription which seemed to have been used as a threshold in the tower. Some blocks of peperino o-6o metre high had been used in its construction : I also saw there part of a gray granite column, the lower part of a female statue, and much ancient debris, both here and to the W.

DlS 0x54 Dis M\_anibus 079 'ARIA Varia\e\ metre S-BRE Bre •\RJV Variu\s\ CO co\niugi] R/S 0-035 \ca\ris\simae~\

To the S.S.E. I have observed some ruins, but a new vineyard has recently been made : further on in the same direction on the edge of the Macchia, two ruins may be noticed—one a concrete wall with buttresses, the other a small underground reservoir. Here, not far from the Macchia delle Sterpare, was found an inscription which seems to render it clear that here, as in the Vigna Gentilini, and near Roma Vecchia {Papers, iv. 125, 142) there was a vicus or village community, the name of which we do not know. The inscription, published by Grossi-Gondi (p. 179) is cut upon a marble slab C65 by C47 metre, and runs thus : Aiulus) Murrins D(ecimi) l{ibertus) Alexsander mag(ister) veici iteriuni) aram dedit Herciuli) opseqiuenti) liibens) m{erito) Piublins) Murrius Piiibli) liibertus) Phileros magiister) restituit. On the back are the letters CN . MG, the meaning of which is not clear to me. The orthography of the inscription would lead us to assign it to the end of the Republic or the beginning of the Empire : we have the same spelling veici in the inscription found near Roma Vecchia. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 297

Tomassetti {loc. cit.) tells us that an inscription ' now belonging to Tusculum' of Q. Caecilius Hilarus was found here.1 Beyond this point the Via Cavona itself presents no special features of interest nor traces of antiquity, and the neighbourhood of Pantano Secco (the ancient Lacus Regillus) may best be dealt with in connexion with the ancient roads leading to it from Frascati. We may therefore follow the path which ascends S.S.E. from the neighbourhood of Casamari along the Colle Pizzuto, and so into the modern highroad. On the W. side of this path, just W. of the southernmost of the ruins mentioned above, are the remains of another narrow reservoir. From this point the path begins to present traces of antiquity, though Grossi-Gondi is no doubt right (p. 109, n. 1, cf. his map) in supposing that it was prolonged as far as the Via Cavona : indeed he mentions (p. 179) that, near the spot where the inscription of the Murrii was found, there were traces of an ancient paved road. Stevenson, however, saw no remains of it when the railway was made, and only notes {Vat. Lat. 10572, 4iv) that in the cutting just to the E. of the crossing there were found large pozzolana quarries, possibly of Roman date. On the Colle Pizzuto (we do not know exactly where) was found the unimportant sepulchral inscription C.I.L. xiv. 2740. That it justifies us in placing here a villa of the Rutilii, as Grossi-Gondi conjectures (p. 174), is most improbable, and his idea that C.I.L. xiv. 2683 was also perhaps found here is contradicted by Nibby, Schede, i. 65 {infra, 341). After crossing the railway the path continues to ascend and there are pavingstones in the fieldwalls. On the E. side is a small reservoir (W. of the house marked Di Nicola) ; and in the wall of the vineyard in which it is, I saw in December, 1897, a bas-relief deeply cut in a block of Pentelic marble, representing a female draped figure some 1*50 metre long and

1 Stevenson notes ( Vat. Lat. 10572, mv) the discovery of a brickstamp with hollow letters at 145 metres above sea-level, near the reservoir S. of Casamari, probably from the W. of the path. It seems to be unpublished, and is as follows : THE BRITISH SCHOOL AT ROME.

C07 metre wide (Wells mentions what may be the head of this relief on P- 97)- On the W. edge of the path is a small structure of selce concrete, about 4 metres square, with a chamber inside it, possibly a tomb, indicated in Stevenson's map. He also marks, W. of the Casalaccio, where are some architectural tufa fragments, the remains of the platform of a villa. The communal boundary line between the Communes of Rome and Frascati has for some time followed our road, and that portion of it going W, to the Via Tuscolana may mark the line of another ancient road : there are pavingstones (not in situ) along its first portion, but these may also have belonged to the road along Colle Pizzuto ; and along the rest there are no traces of antiquity. After crossing the Frascati railway we see the actual pavement of the road in the bank on the edge of the path, running S.E. by S. Some of its pavement, was, according to Wells, excavated in 1874. The bank soon rises in level, so that the pavement is at least 20 feet under the soil; and there is ancient debris visible, but only at a low level, on the N.E. side. We then see some remains of concrete foundations on this same side, and a piece of the pavement in the modern path, running S. by E., with the margines preserved, from which its width can be ascertained to be 2-38 metres (8 Roman feet). After this point it kept S.W. of the modern track instead of coinciding with it (as is wrongly indicated in my map), so that the deep cutting through which one now passes is not of ancient origin after all: there is indeed, about 50 yards before the high- road is reached, a drain of selce concrete running S.E. by E., about two Roman feet wide, in the bank on the N.E. ; and the pavement that is now visible, a metre or two above the level of the path, is mediaeval or modern. We then enter the highroad opposite the Villa Pescatore. We must now return to the Vigne di S. Matteo, where the Via Cavona intersects the highroad, and follow the latter up to the Villa Pescatore and then to Frascati. On the S.W. side of the road, opposite the house at point 153, is a platform of a villa on the edge of the hill : there is only a little preserved in situ, but there is much debris, including the slabs and bases belonging to a peristyle with peperino columns 038 metre in diameter. This is not indicated in the map; but further S. on the E. side of the railway, the remains of yet another villa were found in the new vineyards—we were CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 299 told of walls and a cistern, and of a terracotta figure about 2 feet high, but only saw debris. Stevenson notes the discovery of two capitals in making the railway below this villa not far from the high-road. On the N.E. side of the highroad two rock-cut cuniculi were found, and a little further on, close to the path which crosses the railway E. of point 158, some ancient walls, and between them a level space ten metres wide paved with rough blocks of selce not worn—the pavement either of a lane or of a yard. From the point of intersection of the Via Cavona up to point 183 the Via Tuscolana marks the boundary of the Communes of Rome and Frascati, which then turns E. and runs to the road along the Colle Pizzuto {supra, 297). Just to the N. of" the Frascati railway and W. of Ponte Tusculo, at point 188, are the remains of another villa, where we saw three Corinthian pilaster capitals in peperino, the columns to which they belonged being C42 metre in diameter. To the S. of the line, where the Staff Map marks two separate buildings, entitling them Grotte, there is really only one group. Not very far from here Giorgi must have seen in October, 1734, the sepulchral inscription of A. Fabius Proculus, prefect of the first cohort of Dacians, and tribunus militum legionis II. Adiutricis (C.I.L. xiv. 2618), which he notes as lying on the ground near some substructions in the country not far from the Frascati road, beyond the ascent of Vermicino, on the right going to Frascati. Further to the S.W. is debris, probably belonging to another villa, which was reached by the short branch path running N.E. mentioned in Papers iv. 133. At point 186 a path diverges S.S.E. from the highroad, which is almost certainly of ancient origin, though there are no traces of antiquity upon it, and soon reaches Torre di Micara. Its course beyond that point is described ibid. 134 and supra, 244. Beyond point 183, up to the Villa Borsari, an old boundary wall, marked in the map, runs parallel to and just N.E. of the modern highroad : there are, however, no pavingstones in it, though one might have expected to find them, had it marked the line of the ancient road. Between it and the modern highroad, at the bend S.E. of point 186, are traces of a building, in the vineyard. At the Villa Borsari itself there are no remains of antiquity ; but to the E. of it there is a ruined farmhouse resting on foundations belonging to a building, the main part of which was circular, with another apse facing 300 THE BRITISH SCHOOL AT ROME.

N. constructed of opus reticulatum, and to the E. of that debris of another building. Just S. of the Villa Borsari another path, the prolongation of that from Casale Ciampino to Torre di Micara {Papers, iv. 133), falls into the Via Tuscolana, which now describes a considerable curve to avoid a valley : so that I have conjectured (see the map) that the ancient line passed close to the Villa Borsari and to the buildings I have just mentioned. Thence it ran almost due E. following the line of an older boundary wall across the zigzags, in which pavingstones are visible, while just below the Villa Pescatore, on the S. side of the modern road and some seven feet above its level, a little paving is visible in situ in the bank : the blocks, which are smaller than usual, are O"2i metre thick, and are set in a bed of mortar about O"23 thick, which rests on the solid rock. There are also pavingstones in the fieldwall N. of the road ; but, from the indications I have given, Grossi-Gondi (pp. 105 sqq.: see his map) seems to be wrong in making the ancient road pass N. of the modern at this point. He, it should be noted, does not admit the antiquity of the section of the highroad from Fontanile Vermicino to Villa Borsari, and brings the ancient Via Tuscolana across from Casale Ciampino past Torre di Micara to Villa Borsari and Villa Pescatore, and so into the modern highroad, following Nibby's idea, no doubt {Papers, iv. 133). Fabretti {De Aquis, Diss. I, tab. I, and map opp. p. 90) makes the Via Tuscolana diverge from the Via Latina before Morena and pass N.E. of Centroni (I think without good reason). Mattei, commenting on this (p. 20), remarked that its pavement was to be seen near the vineyard of the Jesuits {infra, 301) and below the walls of Frascati near the Accoramboni palace. At the Villa Pescatore, now Cicinelli, there are no traces of antiquity, though there is a good deal of debris nearJ; but the Villa Sora, constructed by the Monni family and bought from them by Giacomo Boncompagni, Duke of Sora, in 1600, is built on the remains of an ancient villa, some of the vaults and walls of which exist immediately to the S.E, of the house. The pavement of a road was also found, as to the direction of which I could obtain no certain information. The main hall of the Villa Sora itself (now converted into the chapel of a school of the Salesian fathers) contains some very fair paintings by the Zuccari; and the

1 C.I.L. xiv. 2654 has been copied here, but is said to have been found at Ciampino: so has also id. vi. 10004, found outside Porta Maggiore in the sixteenth century. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 301

Director's room contains a good copy of Guercino's Aurora in the Casino dell' Aurora in Rome, and the portraits of two Popes of the Ludovisi family. Above this villa there are pavingstones in the bank on the S.W. side of the road, 3 or 4 feet above the modern level, some of which may be in situ (Grossi-Gondi indeed marks the ancient road as keeping just on this side of the modern, and it may have been part of it that was found within the Villa Sora), while in the wall on the N.E. are pavingstones also ; and at the entrance to the Orti Sora is a theatrical mask in white marble (from a sarcophagus), and there is another at the Casale used as a fountain jet (Tomassetti, 177). The brickstamps C.I.L. xv. 371, a. 19 (period of Severus) 1094 e. 36 (75-100 A.D.) are recorded by Volpi {Vetus Lalium, viii. 155) as found in the vineyard of the Jesuit College of Frascati on the modern Via Tuscolana, beyond the tenth milestone. This vineyard was known as Prete (for Pietre) Liscie, i.e. the pavement of the ancient road was visible in it; and it was bought under Paul III. (Grossi-Gondi, Villa dei Quintilii, p. 4, n. 2). Marini {her. dol. n. 176) records the former from Lupi's notes as found at the Rufinella, but he may be in error. I have not yet been able to apply the test of examining Lupi's own copy in the mass of miscellaneous archaeological notes collected by Marini ; and possibly it is not preserved among them, or it would be noted in the Corpus} At the church of $• Maria di Capo Croce the road is joined by the path from Villa Pallavicini {supra, 246) and soon reaches Frascati. Here, instead of passing simply along the S.W. side of the modern town, as I have indicated, it seems to have turned to the E. (as Grossi-Gondi has more correctly shown, cf. Lanciani, Bull. Com. 1884, tav. xx.) to avoid a valley which was in the main filled up in 1884, when the new railway station was built. Remains of its pavement were found above the church, between it and the Palazzo Micara, in the embankment of the station itself {Not. Scavi, 1884, 348) and to the N. of the Via dei Merli {id. 1885, 478). It then kept outside the mediaeval walls of Frascati until it reached the level of the Porta San Pietro, which was only opened by Innocent X. (1644-55)- As Grossi-Gondi remarks, a fair idea of its course can be obtained from the Vero e novo disegno di Frascati con tutte le ville convicini made by Matteo Greuter in 1620 (a later edition was published by Giovanni

1 C.I.L. xiv. 2759 (a sepulchral inscription) was found in 1852 'between Ciampino and Frascati.' 302 THE BRITISH SCHOOL AT ROME.

Giacomo Rossi) and repeated in three plates in Kircher's Vetus Latium (PP- 77-9)- Frascati was identified with Tusculurn by certain of the topographers of the seventeenth and eighteenth centuries, such as Holste and Mattei; but as this theory has now been definitely proved to be erroneous, we need not concern ourselves further with it. I shall only add that one of the grounds on which they based their theory is an element in the controversy as to the antiquity of the Via Tuscolana, which I omitted to take into account in Papers, iv. 51. I allude to the statement of Dionysius of Halicarnassus (x. 20) that Tusculum was not less than 100 stadia (i2i Roman miles) from Rome. This statement has led some writers to seek for a road more direct than the Via Latina, by which the distance would have been about 14 miles (112 stadia) up to the amphitheatre. Canina {Tusculo, 66), considering the modern Via Tuscolana as ancient, calculated the distance to the Villa Rufinella as about 100 stadia. Lugari, Monutnenti al iv miglio della Via Appia, tav. viii, was of the same opinion, and Tomassetti (who had at first considered the Via Tuscolana as mediaeval) has later, in a pamphlet on the Festa del Tusculo (Rome, 1899, 14) returned to Canina's view. Nibby and Cozza, on the other hand, as we have seen, considered the Via Tuscolana to be the deverticulum which left the Via Latina at the tenth mile {supra, 300). To my mind Dionysius' indication does not pretend to be accurate, and proves nothing one way or the other. The 112 stadia along the Via Latina are naturally measured from the Porta Capena, and it is another mile to the outskirts of Rome at least; so that he might very fairly put the distance at quite 100 stadia. I think Tomassetti (Via Latina, 48) is quite right in calling the strata nova Tusculana of the Bull of Honorius III. the Via Tuscolana, and the strata antiqua Tusculana the Via Latina, but it is going too far to say that it proves that the former is of mediaeval origin. Grossi-Gondi (pp. 105 sqq) gives an excellent summary of the question. Other writers have attributed the ruins upon which the mediaeval town was built to the villa of Lucullus, but the arguments they have brought seem insufficient and, as we have seen (supra, 249), there is some reason for supposing that the Villa Conti (now Torlonia) really occupies the site of the villa of Lucullus. On the other hand Grossi-Gondi (124 sqq.) appears to have grounds for his supposition that these ruins belong to the villa of Passienus Crispus, and I notice that Lanciani (Wanderings in the Roman Campagna, 283) expresses himself as convinced CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 303 by his arguments. The passage from Pliny, which has already been quoted {supra, 252) concerning Crispus' affection for a peculiarly fine tree on the hill of Corne in the territory of Tusculum, seems to show at any rate that he had a villa at Tusculum ; and the discovery in Frascati itself in 1876, under the foundations of the house of the Sturbinetti family (now the property of the Lugari family) of a lead pipe bearing the name of an Agrippina {C.I.L. xv. 7853) and the fact that the younger Agrippina, the mother of Nero, was his second wife, and indeed compassed his death for the sake of his property (Suetonius, frag., p. 290, Roth) goes far to justify the identification. The inscription on the pipe is unluckily fragmentary— u. Agrippinae is all that is preserved, and it is of course possible that, as Lanciani thought, this was another Agrippina altogether— either Asinia or Vipsania Agrippina {Prosopograpkia, i. p. 169, no. 1041 ; iii. p. 443, no. 462) or (Vib)u(llia) Agrippina. as Dressel conjectures {ibid. iii. p. 431, no. 425). But there is further evidence in favour of the identification; and Grossi-Gondi rightly insists that in view of a passage of Tacitus {Ann. xiv. 3) igitur Nero vitare secretos eius congresses, abscedentem in hortos aut Tusculanuin vel Antiatem in agrum landare, quod otinin capesseret, and of the discovery in 1891 (in the contrada Croce Bianca, hardly a hundred metres from the house of the Sturbinetti family) of another lead pipe bearing the inscription Neronis Claudi Caesaris {C.I.L. xv. 7817) we cannot refuse to accept it. Additional evidence may be found in the sepulchral inscription of a freedwoman of one of the Claudian emperors, found, it is said, in Frascati itself in the foundations of the Palazzo Senni—-close, that is, to the ancient road, as Grossi-Gondi (p. 133, n. 2) points out. {C.I.L. xiv. 2690 Dis Manibus Claudiae Primigeniae lib{ertae) Aug{usti) benemerenti Mamilia Albana jiliae vixit annis XV. mensibus iiii.) And we find that the site of Frascati remained imperial property under the Flavian emperors : another lead pipe {id. xv. 7818) was found under the Casa Sturbinetti in 1876, bearing the inscription Imp{eratoris) Domitiani Caes{aris) Aug{usti) sub cura Alypi l{iberti) procurator is) fecit Abascantus ser{vus) Atime{tianus). It is not certain whether one or two copies of this pipe were found under the Casa Sturbinetti : it seems to me probable that two were found and that one of them is to be identified with that of which Henzen speaks as found near Porta Granara (which was close by) in the ruins of an ancient 304 THE BRITISH SCHOOL AT ROME. villa. Mattei tells us indeed (op. cit. 59) that two years before he wrote (1711) the statues of Domitian and Domitia were found, each in its own niche, in the Villa Cremona, while the land belonging to Signor Muzio Massimi was being turned into a vineyard. These statues, he says, were bought by Ficoroni and placed in the gallery of the Duchess Rospigliosi. It does not seem to have occurred to anyone to look for them in the Palazzo Rospigliosi, and I do not find that any writer mentions the provenance of the two statues Matz-Duhn 1343, 1501, which I think must be identified with them ; though in neither case is the attribution certain according to modern canons of iconography. The first is a statue of Titus in the act of addressing his soldiers (adlocutio) (Bernoulli, Rom. Ikon. ii. 2, p. 33, no. 9), probably that mentioned by Winckelmann, Gesch. der Kunst, xi. 3. § 21 = Werke (Donanoschingen, 1825), vi. 244, n. 3). The second is a draped female statue, the head of which has the hair dressed like the so-called Domitia from Herculaneum (see Bernoulli, op. cit. p. 63 : he does not actually mention our statue) but perhaps does not belong to the body. We find, too, a sepulchral inscription set up in honour of his wife Gavia Helpis by a freedman of one of the Flavian emperors, T(itus) Flavins Aug(usti) liipertus) Epaphra procurator) villaruni Tusculanarnni (C.I.L. xiv. 2608). For, besides the villa on which the town of Frascati was built, there were others which were also imperial property—those of Tiberius (Vigna Spinetta) and Galba (Casale Campitelli) and Nerva (?) (Cocciano) (infra, 313 sqq.). Grossi-Gondi further brings together (pp. 133 sqq.) various inscriptions which he connects with the imperial house, which have been found in the territory oi Tusculum : the great majority of these, however, simply belong to persons of the same gens as the various emperors, which, as he recognizes in the case of the Flavii, proves little or nothing, unless it is definitely stated that the person was a freedman of the imperial house; and as he further recognizes, they are, as one might expect, widely scattered, and therefore do not give us much topographical assistance. It has indeed been too often the practice in dealing with the topography of this district to assign a villa to one of the great families of Rome simply because a sepulchral inscription of some quite unimportant person, whether a free man or a slave, has been found in its neighbourhood.1 He rightly notes, on the other hand, that two or 1 I may add, apropos of his footnote p. 134, n. 7, that I do not see why the Pineto dei Sacchetti, where C.I.L. xiv. 852 was found (it is the sepulchral inscription of a freedman of the gens Cocceia) should not be that line of pines situated on the hill above the now ruined Villa Sacchetti CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 305 three other villas, that of Matidia, perhaps another, on the ruins of which is the so-called Barco di Borghese, and that of the Quintilii, became later on parts of the imperial domain {infra, 327, 370, 375). The subsequent history of this domain is decidedly obscure: that it is not actually named as imperial property after the time of Domitian is no doubt simply due to the fact that chance has preserved to us no other inscriptions. It does not occur among the donations made to the various churches of Rome and its neighbourhood by Constantine, which may indicate that it had already passed out of the hands of the imperial house—but to what owners we have no means of knowing. Nor have we any knowledge of the history of Frascati until we find it appearing for the first time in the ninth century j1 for, as Grossi-Gondi has pointed out {Bull, Com. 1906, 30), the legend of the donation of Frascati or of Tusculum itself to S. Benedict in the sixth century A.D. has no real foundation in fact, and all we can consider as certain is that the abbey of Subiaco possessed certain landed property in the territory of Tusculum which had previously belonged to the monastery of S. Erasmus on the Caelian (the fundus Africani and the fundus Oppiniani—Papers, iv. 128), which had come to it in 937. We must now turn to the description of the villa upon which the mediaeval town of Frascati was built. The extent of the latter is clear from Greuter's bird's-eye view, already cited : its S.E. wall corresponds with the street which runs to the N.W. of the present cathedral of S. Pietro, which thus remained outside, the original cathedral being S. Maria di Vivaro, now S. Rocco, a good deal lower down.2 But the villa seems to have extended somewhat further S.E.—according to Lanciani {Bull. Com. 1884, 182) as far as the Villa Lancellotti, according to Grossi-Gondi (p. 122) as far as the back of the cathedral. I saw a probably ancient vaulted chamber in a house on the S.E. side of this street, a little N.E. of the cathedral; and in 1905, a little to the to the W. of Monte Mario. And as to the question of the origin of the name Cocciano {infra, 316) I think he is right in refusing to derive it from Cocceius, the gentile name of Nerva : Tomassetti's conjecture ( Via Latina, 194) that it is derived from Chaucianum, i.e. that this is the Villa of Gabinius, who had defeated the Cfoauci, is even less probable. 1 For the mediaeval history of Frascati I must refer to Tomassetti, Via Latina, 170 sqq. ; Lugari, L'origine di Frascati, Rome, 1891 (also in Diss. Pont. Ace. Arch., Ser. II. vol. vi. (1891)). 2 For the discovery of its remains in or about 1732 cf. Volpi, Vetus Laliunt, viii. 10 and the comments of De Rossi, Bull. Crist. 1872, 159 (French edition). X 306 THE BRITISH SCHOOL AT ROME.

S.W. of the latter, a drain or aqueduct was found about 2 feet below the ground level : it was in selce concrete with a round top, about O'5o metre wide and C90 high. Nibby {Analisi, iii. 340 i^.) and Lanciani {Bull. Com. cit.) describe the other remains formerly visible in the town, the N.W. wall of which followed the limits of the upper terrace—the substruction walls on the N.E. side, and various other remains, some of which are not visible, while the most important of those which are still existing are the two chambers behind the apse of S. Rocco, each 5 metres wide (these are not identical with the reservoirs mentioned by Grossi-Gondi, p. 124, n. 3 : for the springs which fed them cf. ibid. n. 2 and see also Mattei, Tuscolo, 68). In 1858 some unimportant fragments of sculpture were found in a street below the walls (cf. the same reports), while in 1848, in the Piazza del Gesu, a fine Ionic pilaster capital and fragments of friezes were found. {Atti del Camerlengato, Tit. iv. fasc. 3609, cited by Tomassetti, Via Latina 188, cf. ibid. 3710 for the discovery near the church of S. Rocco of an ancient building with brickstamps of Aurelius Rufus {C.I.L. xv. 882—end of second century A.D.), etc.. Tomassetti speaks of discoveries of pieces of pavements, columns, etc., in the cellars of the Casa Petri and of painted rooms and marble fragments in the Via Varardesca : one of these last—a fragment of a group of a wolf and a draped figure—was seen by Wells (p. 71) in the public garden. In all these buildings the concrete is faced with chips of selce in the foundations, and with opus reticulatum of selce with brick quoins above ground. The lower terrace of the villa is not indicated in my map but is marked by Grossi-Gondi as extending from the mediaeval city wall almost as far as the railway on the N.W., right under the monastery of the Riformati: Nibby speaks of a large reservoir under the monastery and Lanciani notes a wall of opus reticulatum with windows and arches 58 metres long under the left hand side of the church : this is no longer visible. Greuter's view seems to show remains even further down, close to the house just below the railway, N.E. of Capo Croce. The passage quoted by Lanciani (p. 184) from Cod. Tusc. 14, i. 11, ff. 146 sqq., must refer to this lower platform. It speaks of remains of the villa visible below the Porta Romana, extending as far as the garden and palace of the Cherubini family, afterwards bought by Colonel Guaina, a citizen of Rome, who found CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 307 there some fine statues, which he transported to his palace in Rome. The place, Mattei {Tuscolo, 68) tells us, bore the name of Bagnara, and terracotta pipes, both round and rectangular, seem to have been found there; and in 1695, in the ruins of a house which stood there, a marble head, resembling that of a sheep, was found, also capitals of columns coated with stucco, and other marbles, with fine pavements ; and it was reported that a treasure had been found there, with valuable marbles, lapislazuli, etc. This is possibly identical with the so-called nymphaeum of Lucullus, below the Villetta Pentini,1 which was discovered in 1854. Lanciani cites Visconti's account in the Archives cf the Pontifical Ministry of Fine Arts, from which it appears that at a depth of 4 metres a very fine pavement of white mosaic was found, and also a column drum of rare breccia traccagnina. Wells (p. 74) mentions statuettes found here and two busts. At Frascati in the Piazza Baldassare Peruzzi drew {Uffizi, 416) a plain Ionic column base—the size at the bottom was p. 4 oncie 11, i.e. about ro8 metre. Antonio da Sangallo the younger {Uffizi, 1184) has a drawing of a base at Frascati. Dosio (z'£. 2011 v) has a drawing of an Ionic base from a building ' detto di Lucullo.' There are in the public garden some unimportant antiques of uncertain provenance, noted by Tomassetti {op. cit. 188, note). A fine fluted column drum of porta santa marble found in repairing a drain under the Via dei Merli was also placed here {Not. Scavi, 1884, 348). Tomassetti also gives {Bull. Com. 1887, 239) a dedication to Silvanus, bearing the date Jan. 5th, 179, and {Not. Scavi, 1895, 350) a number of fragmentary inscriptions, most of them preserved at the Episcopal seminary, and mentions a sarcophagus in the courtyard of the bishop's palace (Via Latina, p. 255, n. 1). Some unimportant sepulchral inscriptions in the Municipal collection at Frascati, and others in the pavement of the Cathedral are published by him in Bull. Com. 1895, 162, and Not. Scavi, 1895, 351. Wells (p. 258) notes various fragments of sculpture in the town; and a fragmentary 1 Near the garden of the Villetta Pentini was found in 1895 a fragment of an inscription with an elogium of M. Vinicius, consul in 19 B.C. As it is on a small scale—the letters of the first two lines are CTO5 metre high, the rest

O/25 m.

In 1883-4 Lanciani saw various antiquities in the possession of Alessandro Fausti, a dealer in antiquities, including C.I.L. xv. 1443, 2266, found at Fontana Candida, also a fragment of a fine marble tazza, a portrait head of a boy, perhaps of the imperial family, the capital of a pilaster o-68 metre high and 070 wide, etc. {Not. Scavi, 1883, 85 ; 1884, 83). In 1886 he saw lead pipes in Fausti's possession with the inscriptions C.I.L. xv. 7831, 2. Atriae Moscharus c{larissimae) f{eminae)—one example of which has since (in 1892) been found in the tenuta di Pantano —cf. Papers, i. 198—786S b 2. T. Statili{us) Felicio fedit)—o{ which one example was found at Gabii in 1792, and another in the tenuta di Pantano in 1892. Helbig copied in Fausti's possession the unimportant dedication C.I.L. xiv. 2587 a. Dressel also publishes some lead pipes ' found in the territory of Frascati' as to which we know no other details—C.I.L. xi. 7839 (?) N, Iuni Fausti, 7850 (?), C. Valeri Paulini, 7867, 2, Septimius Secundinus fecit (cf. infra, 393), 7870 C. Vettenius Felix fec{if) (as to which cf. Stevenson, Vat. Lat. 10572, 3V) 7876 (?). A mosaic formerly in the Villa Cavalieri at Frascati, where it was seen by Montfaucon, Ant. Expl. (Suppl. ii. pi. 23, cf. Furietti, De Musivis, 58) was republished by Guattani {Meniorie Enciclopediche, iii. 45, sea.), in whose time it was in the possession of the mosaic maker Volpini: Guattani interpreted it as representing the question put by Zeus and Hera to Teiresias, as to whether the male or the female had more pleasure in love. In 1880 it was in the possession of one Scalambrini, a dealer in antiquities in Rome, but has since disappeared. Engeimann {Rom. Mitt. xix. (1904) 286) interpreted it as a representation of the quarrel between Erechtheus and Eumolpus. At the Casino Marconi, opposite to the E. entrance to the Villa Conti, CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 309

according to a statement of Fea {Nnova descrizione dei monumenti antichi (Rome, 1819), 87), there was in his time a complete replica, placed more expressively upon a rock, of the famous headless statue of a daughter of Niobe in the Museo Chiaramonti in the Vatican (no. 176). Of this replica, however, no traces are to be found. A description of the Marconi collection will be found in Guattani's Memorie Encidopediche, iv. 1 sqq. He mentions a Mars (now in the Lateran, no. 127) which was said to have come from the collection of Gavin Hamilton, standing on the cinerary urn with the inscription C.I.L. vi. 10958 (which was found near S. Sebastiano on the Via Appia in 1793), a Bacchante, an ideal bust said to have come from Acqua Traversa, an Aesculapius (Lateran 182), a head of Pallas, a statue of an empress, a bust of a Flavian emperor, a bust of Juno, a statue of Euterpe (Lateran 187) a bust of a Faun, a statue of Diana, and a bust of Annius Verus. The provenance of these is by no means certain, and they may not have come from Frascati. Wells (Alban Hills, 38, cf. 115) mentions other statues, which probably came from the Roman collection of Marchese Campana (to whom the Marconi collection then belonged), inasmuch as she copied there the inscription I.G. xiv. 1109, which was found in Rome in 1660, and was seen in the nineteenth century by Teza at Florence in the possession of a dealer ; and she mentions that Campana lived there (Kaibel was not aware that Wells had copied it here). De Rossi, Inscr. Christ. 1060, gives a fragmentary inscription from Mattei, Tusculo, 93, con vix ann. xv. inensi. .. .p^\ c. Vilisari... ., which was found near the ' tomb of Nero ' by Giuseppe Catani, a sculptor, in very fine lettering, p.c. stand for post consulatum, so that the date is 537 or the beginning of 538 A.D., during the siege of the city by the Goths ; and it indicates, De Rossi thinks, that Belisarius held Tusculum at the time, as he held Tibur. But neither De Rossi nor Dessau has noticed that Mattei gives immediately under this inscription, the following, Sex. Naevius Sex. 1. Philemo Naevia Sex. I. Tuscula in f. p. xv. in ag. p. xix., which he ascribes to Gruter, Inscripliones, p. 986. This is given in C.I.L. xiv. 3363, with the reading Arbuscula, from Fabricius, In Horatium argumenta et castigationes (Leipzig, 1571), p. 89, but is there rightly ascribed to Praeneste. Mattei gives it without any comment, but, as he calls the first inscription ' another fragment like' that on which are cut C.I.L. xiv. 2765, 2766 ( = De Rossi, cit. 1061) (one on one side, one on the other), it 310 THE BRITISH SCHOOL AT ROME. may be that he would wish us to believe that it was cut on the reverse of the stone on which was C.I.L. xiv. 2363 ; but in that case we may not without reason be a little suspicious of its authenticity, or at any rate of its belonging to Frascati. On the reverse of the leaf in Stevenson's notes {Vat. Lat. 10572, 143) but in De Rossi's (?) handwriting, on which are given the Christian inscriptions, 1060, 1061, and four others from the cemetery of S. Zoticus {Papers, i. 242), I find some other inscriptions which do not seem to have been published : {a) Part of a relief ; three men lying on a triclinium, with bread and fish before them :—

(b) Sarcophagus with strigil moulding; on one side a shepherd leaning on a staff, on the other a veiled woman with her head raised ; in the centre two lambs holding a clipeus with the inscription :— IANVA RI IN PA CE DOM {c) Inscription—no details :— VONBIA SATURNINA HIC DORMIT PATRONE VE NEMERENTI {d~) Fragment:— ^ATIS IN JoCTOB A tomb was found in 1886 near Frascati with the skeleton still preserved and a bronze collar round the neck, with the inscription : tejie me et reboca me Aproniano Palatino ad mappa aurea in Abentino quia fugi [Bull. Com. 1887, 265, 289 : C.I.L. xv. 7182). The name mappa aurea is also mentioned in the Notitia, but whether it is that of a building or of a street is uncertain (Jordan-Hlilsen, Topograpkie, i. 3. 170). CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 311

There are a number of discoveries of which it is recorded simply that they were made at or near Frascati, further details not being given. I have thought it better simply to group them here. Volpi( Vetus Latium, according to Muratori, 1645, 8) gives a sepulchral inscription which begins thus Bao-cro? eywv oh' eKelvot ov efcrave Av]opo<; avrjp (Kaibel, ib. xiv. 1505) as found near Frascati. A bronze bell found near Frascati is now in the Museo Kircheriano : it bears the inscription JLicrcnreoov TlpcoToyevi] vUa, Isapio being the name of a charioteer apparently (cf. Kaibel, I.G. xiv. 2409, 2. C.I.L. xv. 7233, Riccy, Alba Longa, no). The sepulchral inscriptions found or copied at Frascati or in its buildings without any accurate record of their provenance are C.I.L. xiv 2672, 2674-2676, 2692 ( a tomb of various people on a site granted by Q. Pompeius Falco, the friend of Pliny, cf. Papers, i. 237), 2696, 2709 (a metrical epitaph set up by M. Gellius Maximus to his freedman Phoebus), 2717 (a cippus erected in memory of a boy, C. Iulius Saecularis, with his portrait above), 2725, 2726. In the Vigna Buzzi was found, according to Fea's notes, the dolium stamp C.I.L. xv. 2257. Eschinardi • (Esposizione cit. 369, 387) speaks of a detailed topo- graphical map of the territory of Frascati and other neighbouring places as far as Nemi, etc., printed by himself not many years before. This map is unknown to me. Still more interesting would be the map of the Campagna di Roma printed in the year 1513 in the pontificate of Leo X., of which Mattel (p. 39) speaks, and which Lanciani (Storm degli Scavi, iii. 54) has never seen. Tomassetti, Campagna Romana, i. 247 (Rome, 1910) does not mention it, and states, so far as I know, correctly, that the earliest existing map of the Campagna is that of 1547. Aldovrandi (pp. 150, 151, 158) notes the existence in the Palazzo Farnese of a fine trophy decorated with a Medusa in the centre, of a trophy of porphyry, and of a triangular candelabrum with winged Victories in relief, a winged Roma triumphans on the side and harpies at the feet, all found at Frascati. The following sculptures in the British Museum were found at or near Frascati (no further details being known): all come from the Townley collection : a terminal figure of a youth in the character of Hermes, found in 1772, MS. (Catalogue of Greek Sculpture, no. 1605), the head of a Muse 3i2 THE BRITISH SCHOOL AT ROME.

(found by Gavin Hamilton, no. 1691), the front of a sarcophagus, with Cupids in the circus (no. 2319), a console (no. 2608). The Museum of Fine Arts in Boston acquired in 1899 a statuette of an apoxyomenos, found at Frascati and described by P. Hartwig in Berl. Phil. Wockenschrift, 1897, 3° (See Jahrb. d. lust. xvi. (1901) Anz. 165); and the head of Diomedes at Ny-Carlsberg (no. 147) was found at Frascati, but we have no further details. Having completed the description of Frascati itself, it will now be necessary to deal with the district to the N. of it, between it and the Via Cavona. We may first of all descend by the path which leaves Frascati on the N.N.W. side, passing close by the monastery of the Riformati,1 and crosses the railway almost immediately. It has no traces of antiquity any- where along its course. On the E. side of it just below the railway is a drain and on the W. side a very large reservoir, with a smaller one, partly filled in, adjoining it, in the so-called Orti Sora, a plan of which is given by Uggeri, Giornata Tuscolana, pi. vii = Angelini and Fea, Via Latina, pi. vi, while Canina, Tuscolo, tav. 9. gives a plan of the larger one only. Uggeri also gives a view of the interior ( Vedute, no. 9). The larger reservoir of the two is square: it measures about 36'2O metres each way inside (Lanciani's figure, 4i'28 metres, seems excessive even as a measurement over all—Bull. Corn. 1884, 180). There are six arcades each way, 4-90 metres in width ; the 25 pillars supporting the quadripartite vaulting of the roof are each 1-20 metre square, the corresponding pilasters each project C40 metre from the internal walls, and the exterior wall is C90 metre thick. Canina is right in showing buttresses only on the S. side and not all round, for on the other sides they cannot be seen, the building being below ground level. Monsignor Vespignani, to whom the villa had previously belonged, erected on it the following inscription (Mattei, Tuscolo, 92) Curioso antiquitatis studio receptaculum aquarum ad Tusculanas olim terrnas introspicitur anno dirae luis 1656. The allusion is to a plague which is said to have decimated the inhabitants of Marino (Tomassetti, Via Latina, 129). Of the other reservoir which adjoins it only a part can be seen : Lanciani gives the length as 4i'6o metres (unless this length rests on independent observations, it too is excessive) and the width as 13-80

1 Giorgi saw a copy of the brickstamp C.I.L. xv. 904 (f. 26) (Trajan) in their possession—the provenance is of course uncertain. In the Villa Simonetti close by, Wells saw two busts found in (he Orti Sora (p. 80). CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 313 metres : the four aisles are divided not by pillars, but by walls, pierced by apertures 1 "8o metre apart, and about the same in width. Uggeri indicates the apertures as occurring opposite one another only in alternate rows, as in the Sette Sale at Rome, and shows at least seven aisles ; but this number may be excessive, for only four are shown as actually visible. A little lower down is the Casale Campitelli (so called because it belonged to S. Maria in Campitelli in Rome), which is built upon vaulted substructions, with remains of marble and mosaic pavements, and of opus retiadatum walling, resting upon them : these belong, Lanciani thinks, to the main palace of a villa, which he attributes to the emperor Galba on the ground of the discovery of a lead pipe, given by Mattei, Tuscolo, 89, as found in 1705 in a vineyard close by and bearing the inscription felix ar. imp. ser. Galba V.C.1 As it stands, of course, this inscription is obviously impossible, but Dessau (C.I.L. xiv. 213*), Lanciani, and Grossi-Gondi all consider that it is a corruption of a genuine one. Whether, as Dessau thinks, the legend has been altered to suit the fact that Suetonius mentions {Galba, 4, 18) a villa in the territory of Tusculum belonging to this emperor, or whether a genuine inscription of Galba has been wrongly copied, is a different question. It is equally impossible to say how far the villa extended— that it included the localities called La Sterpara and Pantano Secco, as Tornassetti thinks (Via Latina, 177 n.), is quite unlikely. Lanciani and Grossi-Gondi, whether rightly I rather doubt, both treat the remains of a villa on the so-called Colle Fiorano, to the N.W. of the Casale Campitelli and of the railway, as a part of the upper one. I think, however, that Mattei's account of the discovery of the lead pipe refers rather to the neighbourhood of this villa than of the upper one : for he says that the large reservoirs served for water which ' went by a lead pipe to the nymphaea which lay below the Villa Campitelli at the entrance to the Macchia della Sterpara.... as we know from a piece of a lead pipe found in 1705 in the vineyard of a native of Frascati, where are the aforesaid baths in four chambers, each 80 palms (l7'84 metres) long and 25 (S"57 metres) high.' I do not know, however, what are the baths (' terme') (really, no doubt, a

1 Mattei mentions the existence at the Casale of some well executed bas-reliefs and of C.I.L. xiv. 2765-6 (cf. supra, 309). It seems most likely that the brickstamp C.I.L. xv. 1370. 1 {159-164 A.D.) belongs to these ruins, cf. Wells, p. 88. THE BRITISH SCHOOL AT ROME. « reservoir) to which he refers : the description does not correspond with those of which we have spoken, which are not in a vineyard at all. The altar with the inscription Ara Solis now in the Vatican [C.I.L. xiv. 2583) was found in the upper part of the same vineyard as the pipe. The extensive ruins on the Colle Fiorano (cf. Wells, p. 92) were supported by massive substructions of opus reticulatum with quoins of selce. A cryptoporticus, which was discovered in 1879, has the ceiling of very fine white stucco, coffered, but has now been filled in again. Lanciani notes that he saw columns of sperone, and brick columns coated with painted stucco, and in excavating for the railway just above it in 1883 two fragments of a fine terracotta frieze were found, one with archaic masks, the other with foliage and volutes {Not. Scavi, 1883, 173). Further on is a platform, belonging no doubt to the same villa, which extends to the firing point of the rifle range (Tiro a Segno) : there are substructions belonging to it N.W. of the Casa di Nicola, and at the W. angle over the steep slope down to the Valle Lupara. On the top of the platform there is but little debris. That this was a distinct villa is rendered probable by the fact that it received its water supply from a large reservoir to the E. of the upper platform and of the railway, in an oliveyard which occupies the triangle between the two paths. I give a plan of it here, as it seems to be unknown. I did not mark it in my map, as I owe my knowledge of it to Stevenson's maps. There are five halls with barrel vaults, intercommunicating by arched openings 2*20 metres in span ; it is almost entirely sunk below the ground, which explains the comparative thinness of the outer walls. The total internal measurement is 35*10 by 23^40 metres. I should be inclined to place here the site of the discovery of two statues of which Mattei speaks thus ' I am sure that this villa (that of Galba) extended for a long way where the Macchia now is: for a few years ago our Commune (to whom it belonged) made excavations there, and found a statue without a bust, draped in the heroic manner . . . and another also headless, with a toga... these two statues are now to be seen at Frascati on the stairs of the Palace of the Conservators.' Canina, Tuscolo, 101 n., saw them there, but remarks that they were much damaged. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 315 .» Mattei also records as found near the Villa of Galba the sepulchral inscription C.I.L. xiv. 2691, which belongs to the Republican period. He saw it in the Villa Pescatore {supra, 300). Following the right hand of the two paths across the railway we find no traces of antiquity along it. Stevenson noticed, on the edge of the Macchia della Sterpara (Vat. Lat. 10572, 33V), a place where the ground had been damp and therefore had been drained in ancient times. We continue to follow the path across the railway line, and at point 171, at the second 'a' of 'Macchia' reach a large villa with two platforms, which extended right across the path : the upper platform had shallow arcades in front of it, with a span no less than 670 metres wide in one case : the back wall is faced with very large blocks of opus incertum of selce, one piece measuring as much as C53 by O'4i metre (see PL XXXV. Fig. 2, from a photograph kindly lent me by Mr. J. H. Ten Eyck Burr). The lower platform was supported by a natural cliff of selce, commanding a fine view over a deep valley, the sides of which are clothed with scrub, as the wood is periodically cut for charcoal burning. There is nothing further to notice between this point and the Via Cavona, except scanty debris on the N. extremity of the Colle Fumone, and we may now return to Frascati, and follow the path which leaves it on the N. and runs due N. past the Campo Santo, past the Colle Spinetta, and then divides into two branches, both of which seem to be ancient, and run through the Quarto Cisternole down to Pantano Secco and the Via Cavona.1 On the W. of it is the Villa Sansoni, some antique sculptures in which are described by Wells (p. 256). A little to the E. of it is a house called the Palazzetto (at point 301 in the map) belonging to Signor Mastrofini : in two niches in the external walls are two heads, both on modern busts : on the N.W. side is a female head, Praxitelean in style, and on the S.E. a rather feeble portrait head of a Roman lady ; while Tomassetti (p. 178) mentions four other heads within the villa, two of them of Diana. In 1883-4 in the course of the railway works some brickstamps were found {C.I.L. xv. 315, 12 (Hadrian) 817, 5 (first century A.D.), 824, 6 (early first century A.D.), 1121 a. 7 (first century), 1138, 2, 1489 a. 1 (first century), 2238, 3 (first century)—Not. Scavi, 1884, 83). There is practically no doubt that 807,3 (Hadrian) and 1053,7(135 A.D.) were found on the same site (cf. Not. Scav. 1883, 173, where a jasper of 12 mm. in 1 The first portion of the road is shown somewhat differently by Grossi-Gondi. 316 . THE BRITISH SCHOOL AT ROME. diameter, with the letters IACJ, is mentioned. These are stated to have been found among some walls of opus reticulatum and brick, at the point nearest to the villa of the Quintilii (then identified with the Barco di Borghese). The description ibid. 1883, ^5 of the discovery of walls of opus reticulatum close to the cemetery,; with a much injured Ionic column, and eight very rough amphorae, O'5O metre high, should probably be referred to the same site. Stevenson notes that he saw some debris to the N.W. of the cemetery, and that the path passing by it cut through two ancient walls ( Vat. Lat. 10572, 43,48), with a drain between them, at its N.W. angle. The pavement of the road was found in making the railway in 1884 a little to the N. of the Camposanto {Not. Scavi, 1884, 348). It is also noted that a number of drains hewn in the tufa were found in these cuttings, one of which is still visible a little further S.W. than I have marked it. Here I have indicated (erroneously) a branch road descending from Mondragone and joining our road on the E.: the mistake is due to a misunderstanding on my part of information given me verbally by Grossi- Gondi {infra, 330). Rosa marks a villa here in his unpublished map of the Campagna on a scale of 1:20,000 (now preserved at the Soprintendenza degli Scavi in Rome): there are, however, no ruins now visible on the site. On the W. another path diverges, which may be of ancient origin, though this is uncertain : the ruins at point 226 along it are mediaeval. To the E. in the Quarto Cocciano1 on the Colle Spinetta are the remains of a villa : there is debris in the oliveyard above the modern path, though no walls are visible. The upper platform extends as far as the house at point 281 ; near it are arched substructions (not a cryptoporticus, as Grossi-Gondi says, for there are no windows) faced with rough pieces of selce and some brickwork : here are columns of white marble and tufa. On the lower terrace S.W. of the house is a large open round tank, in the middle of which a spring now rises. The front of the lower terrace is supported on the N.E. by a concrete wall faced with opus reticulatum of selce, and on the N.W. by shallow chambers, originally some 36 in number (one, which I measured, is 4-23 metres deep and 4^40 metres wide and closed in front). To the south of this in a vineyard, Stevenson ( Vat. Lat. 10572, 1) found in 1891 two brickstamps, one circular with the letters . . . ANTI . . . the other rectangular, with the letters .... IERONIS. The latter is possibly 1 That the name Cocciano has any connexion with Cocceius is unlikely {supra, 304 n.). CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 317 a fragment of C.I.L. xv. 2236. Giorgi (ib. 1479) gives a copy of a piece of a brickstamp which he saw built into the house at the Vigna Spinetta (then belonging to one Giulio Balzani of Frascati) which may be a fragment of ib. 127 (126 A.D.). He also copied here the unimportant sepulchral inscrip- tion C.I.L. xiv. 2718. Here, in the Vigna Campoli, there was found in 1892 a lead pipe bearing the inscription (T)i(beri) Caesaris et Iuliae Augnstae {C.I.L. xv. 7814). This seems to be sufficient evidence to make us consider this the villa of Tiberius, and Grossi-Gondi's efforts to prove (p. 170) that Josephus' statement (Antiq. Iud. 18. 170), that his villa was 'about 100 stadia from Rome' suits this site better than Dionysius' measure of' at least IOO stadia ' are, it seems to me, misdirected : neither author is aiming at accuracy, and both are simply giving a round number. The date of the pipe would, as Grossi-Gondi rightly remarks, fall between 9 B.C., when Tiberius married Julia, and 1 B.C., when she was banished to Pandataria. In this connexion it is worth noting that in 1736 (?) a bronze disk (CHX) metre in diameter) was found at Frascati bearing the following inscription on the obverse Thoantis Ti(beri) Caesaris Aug(usti) dispensatoris ab tobis (tori's ?) = couches (C.I.L. xv. 7142 = xiv. 4120. 3). That on the reverse de statione. . . . Caesaris Ang(usti) tabellaris diplomari discede is less clear. Another pipe was found in the Vigna Spinetta (which was on the same level) in 1879, bearing the inscription Imp(eratoris) Vespasiani Aiigiusti) (C.I.L. xv. 7272 : Dressel is, however, wrong in not recognizing the provenance) : so that it is clear that this is one of the imperial villas mentioned above. At the same time two burials under tiles were found ; some of the tiles bore the stamps z$. 1242. a. 4 (first cent. A.D.) 1365. 8 (134 A.D.), and a stamp of 123 A.D. was also found. With one of the bodies was a coin of Trajan. (Cf. Cron. Annellini, 1879, 47.1) Mattei, Tuscolo, 86 sqq., speaks of various discoveries in this neighbourhood : ' in a place called Cocciano, while a certain hillock was being reduced to cultivation, near the road leading to the Torretta (i.e. a tomb which I have not marked, at about the "to" of Q(uar)to S. Marco—Grossi-Gondi, 156: cf. infra, 326) some furnaces were. found, with pipes made of tiles well cemented and nailed together, by which the heat was divided between the upper and lower

1 In this article Lugari mentions a tradition that a pavement of small pieces of marble had been found, and had passed to the Borghese family. 318 THE BRITISH SCHOOL AT ROME. rooms. (I should place the site of this discovery somewhere in the Vigne S. Eusebio.) Close by in the vineyard of the aforesaid Camaldolese fathers, in a place called Spinetta . . . are many ruins, and an enclosure in the shape of a piscina (no doubt the round tank mentioned) : here have been excavated various bas-reliefs, one of them with three faces and a hole down the length in the centre (this must be the base of a candelabrum). In a vineyard a little below, on the N., I had the following altar excavated (C.I.L. xiv. 2755—an unimportant sepulchral inscription), which I said at the beginning was found where are those great walls of stones called "quadrucci" from their shape. In this place is a paved road ten palms wide ... In the same place was also found the following inscription {ibid. 2716. Curiously enough this and the preceding inscription have been seen only once before or since—when Mommsen in or about the year 1845 copied them in the shop of Basseggi, a Roman dealer in antiquities). In another vineyard not far off, where are many remains of ancient buildings, there was found in 1682 the pavement of an ancient temple, all worked in mosaic : in the centre was an altar, and in the corners four rams' heads of bronze, not very large, which I bought from the owner of the vineyard ; but because I was then young, I did not take note by whom it had been erected, nor to what deity the temple was dedicated ; nor did I preserve the inscriptions that were there, which included a tile 3 palms (0*67) square, with the stamp' C.I.L. xv. 1009 (a little later than 108 A.D.). He then goes on to the tenuta of S. Croce to the N.E. {infra, 325). The paved road, 10 palms (2*23 metres) wide, is probably the path to Cisternole. There are, however, various other ancient roads to be noted— one on the S.S.W. ascending steeply from the W.N.W., visible in the bank just below the villa and above the railway, perhaps another going (as Stevenson marked it) along the line of the path parallel to the S.E. side of the villa, and a third possibly following the line of the path below and parallel to this platform, which goes into the Vicolo di Prata Porci ; though neither of these last two has paving in situ, there being only pavingstones in the fieldwalls. The case is similar with a path going N. (not W. as marked in the map) just above the 'C in Colle Spinetta, on the W. of which are ruins, partly of opus reticulatum, including an underground reservoir with two chambers. It is, again, conceivable that many of the pavingstones come CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 319 from the ancient road which ran northwards to the Quarto Cisternole and possibly did not follow precisely the line of the present path, on which indeed traces are rather scanty as far as the fork, though just beyond it I have seen pavement in situ on both branches. At about 225 metres above sea-level the path divides, and a little lower down, in the fork between the two, close to the eastern branch, at point 212, is a large reservoir known as Le Cisternole, which has seven aisles in each direction. A view of the interior is given on PI. XXVI. Fig 1 : since this was taken, the building has been to some extent walled up. The discovery here, in March 1891, of an urn bearing the inscription of a freed man of the gens Vitellia, Q. Vitellius Alexis (Seghetti, Tuscolo e Frascati, 82, n.) gives us no right to suppose either that this family possessed a villa here, or that C.I.L. xiv. 2758 was found here also: Grandi indeed states that it was found with ibid. 2748 below Fontana Candida {infra, 385). In the district of Le Cisternole, in the property of Antonio Benedetti, was found according to Grandi, the Greek sepulchral inscription Kaibel, I.G. xiv. i860. It is the tombstone of a comic actor, and the inscription, which is in hexameters, is commented upon in Bull. Arch. 1873, 49, by Kaibel. The same is the case with C.I.L. xiv. 2686, 2687, 2710, 2724, 2728, 2736, 2751, 2759 a, unimportant sepulchral inscriptions; but Dessau expresses some doubts as to the trustworthiness of Grandi. A little below the cistern there is some pavement in situ ; and a little after the fork Stevenson ( Vat. Lat. 10572, 48) saw a large white marble frieze lying on the ground. There are, too, pavingstones in the fieldwalls. The cutting of the new main railway line to Naples did not, according to Stevenson's notes {ib. 41v sqq.), produce much that was decisive. Beginning from the railwayman's' house (C(asel)lo) we first come to a tomb of concrete, originally faced with rectangular blocks of peperino, measuring externally 9"2O by io'4o metres : the lower part of the interior is cut in the rock, the upper built of huge masses of peperino: there are two shallow niches with flat arches. This must be the tomb mentioned by Nibby (Analisi, iii. 9) and by Wells (p. 275): the latter mentions the discovery, in the early 'sixties, of two stone coffins. No traces of an ancient road were found on the path across the Quarti Perazzeta, which I have doubtfully marked as ancient ; nor does it seem that any traces of antiquity were actually found in situ on the line of either of the two paths from the Quarto Cister- 320 THE BRITISH SCHOOL AT ROME. nole, though there were large pavingstones in the fieldwall on the right of the western part, going down. Instead of that, in making the cutting between these two paths, close to the house at point 192 (the house, he notes,, should be S. and not N. of the line) walls of opus reticulatum with floors, drains, etc. were found, and also the pavement of an ancient road, the stones of which were not at all worn. From the railway we may first follow the western branch which passes along the extremity of a narrow ridge : at point 186 there is a reservoir, the only chamber of which that is accessible is (as far as one can follow it) 52 paces long, and only 1-84 metre wide ; there is another chamber oil the W., filled up, accessible by an opening O'6o metre wide. To the W. of it, on the further side of the path from the Quarti Perazzeta, is another reservoir with two small chambers, which supplied a large villa with two platforms (marked in the map too far apart, as though they were separate). Of these remains Nibby speaks in Analisi, iii. 8 sqq., and he gives an account and a plan, as well as a general plan of Pantano Secco (Sckede, i. 77, 104, 106, 107), but they are not of sufficient interest to merit reproduction. The name of the district was Cornufelle, which some, e.g. Volpi, Vetus Latiuin viii. 172, have connected with Corne {supra, 252) and have desired to place here the Villa of Passienus Crispus, of which, however, we now know the site {supra, 302); others, e.g. Nibby, have connected the name with the gens Corniftcia. Of the upper platform the N. and W. walls in opus incertum are still visible : on the W. is the entrance to a small, vaulted chamber, possibly a corridor ; and further on is a small reservoir in the platform itself; while on the upper level, near the larger reservoir already noticed, is a small passage or niche. The lower platform is supported by curved niches faced with opus incertum of selce with weep-holes, the N. side being continued to the E. by a simple wall in opus reticulatum. Nibby notices that in demolishing those on the W. side it was observed that there was behind them an inclined plane, cemented, descending from S. to N., and above it rubbish to level the soil, from which he infers that the villa originally ended further S., and that when it was enlarged the niches were placed where there was less earth already and the simple wall where there was more. Just below the N. substruction wall there was found in 1822 a CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 321 fragment of a Greek inscription with a dedication to Dionysus, the sun and the moon, which Nibby saw in the possession of one Moroni, who was his guide on this occasion.

AIONYCO) BOIICO HAICO KAI C6AHNH .... KAI BOI Nibby considers that it may date from the 2nd century of the Empire. Wells (p. 276) speaks of the discovery of five bodies in terracotta sarcophagi, of water pipes of lead and terracotta, of coins, etc. To the E. of the villa, the pavement of the ancient road is still visible in the modern path. We now have a good view of the extinct crater basin of Pantano Secco: that it represents the ancient Lake Regillus I have attempted to prove in Rendiconti dei Lincei, vol. vii. (1898) 103 sgg.) and have given a brief summary of the article in Classical Review, xii. (1898) 470. The main arguments to my mind are (1) that this is the only site of those proposed, except Prata Porci, which can safely be said to be in agro Tusculano, and (2) that the latter was certainly not a lake in Roman times (Papers, i. 244), so that we seem by exclusion to be left with Pantano Secco as the only possible representative. It is interesting to notice that the Aqua Claudia and Anio Novus passed, as shown in the map, through the basin, at some height above its floor (cf. Classical Review, xiv. (1900) 327). Nibby (Schede cit., cf. Analisi, iii. 9) is inclined to attribute the emissarium roughly to the 8th century A.D. Just N. of the aqueducts are traces of other remains: Nibby notes here (Schede, 104) 'uncertain ruin, possibly a tomb, near which is a sarcophagus broken and turned upside down.' There are no further traces of the antiquity of the path beyond this point as far as the Via Cavona. There is an ancient reservoir under Casale Marchese, and on the E. side of the Fosso del Cavaliere, and S. of Casale Marchese, low down in the valley, is a rectangular platform supported by walling of rough stones on the N.W. and S.W. It is conceivable that it is the platform of the altar erected to celebrate the victory of the Lake Regillus—for Dionysius (vi. 14) tells us of the sacrifices which occurred on the day after {Rendiconti cit., 124), but this is 322 THE BRITISH SCHOOL AT ROME. very doubtful. Similar walling of rough stones has been found near Gabii {Papers, i. 196) and recent excavations have proved its connexion with a Roman farmhouse. Along the ancient road on the E. side of Pantano Secco (which Grossi- Gondi marks in his map, not considering apparently that on the W. as ancient)1 Stevenson {Vat. Lat. 10572, 48V) noticed pavingstones : S.E. of point 174 the field wall was full of debris, among it a fragment of a large marble block with a few letters of an inscription in fine letters, originally about o-i metre high : ^L^J_fp\ • E. of point 174 was an area enclosed by large squared blocks of peperino, and then a finely preserved piece of the ancient road, a little to the left of the modern path. At Pantano Secco, in the property of Anastasio Reali (which Grossi- Gondi in his map places at point 135) in 1889 was found a sepulchral cippus of L. Septimius Agrippinus, published by Lugari in Cronichetta Armellini, 1889, 100, Tomassetti in Bull. Com. 1895, 161, and Grossi-Gondi {op. cit. 180). The text runs thus: D.M.L. Sep. Agrippinus L. Sep. Antoni Agathonici Nepos v{iri) p{erfectissimi) a rat{ione) h{ereditatiuni) L. Sep. Agathonici v{iri) e{gregii) filius memoriam vivus mihi meisque feci libertis libertabusque posterisque eorum cum loco qui est post dorso memoriae finibus suis una cum casa et aedificio superposito at custodem loci cibariorum gratia pertinebunt. The last word has been added later, and the grammar is clearly incorrect. With it were found ruins of the third century A.D., perhaps of the casa et aedificium itself. The inscription is now in the municipal collection at Frascati. A little to the W. is a small reservoir below the ground level. Stevenson notes traces of the prolongation of the road as far as the Via Cavona past the so-called Grotte Dama, and across it to the Via Labicana, cf. Papers, i. 240, and map iii. and the addendum in iii. 207. Grossi-Gondi (p. 157) rightly notices that it is marked by Fabretti {De Aquis, ed. 1788, Diss. i. tab. i) as Viatrium a Tusculo in Labicanam, though the latter does not indicate its course to the S. of the tunnel by which the Fosso della Morte leaves Prata Porci {Papers, i. 243). We may return once more to Frascati by the Fontanile Trasanella, and then take the Vicolo di Prata Porci. To the E. of the Fontanile are 1 He prolongs the so-called Stradone di S. Marco (infra, 325) hypothetically so as to join this eastern branch just N. of the railway near point 178 (cf. p. 156) y but there, I think, he is wrong, and his error arises from the fact that he does not recognize as even probably ancient the road from Prata Porci to Osteria del Finocchio. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 323 the remains of a villa : further E. are other ruins, for which see Papers, i. 243. Nibby (Sckede, i. 78, 113) noticed the existence (cf. Analisi, iii. 10) in the basin of Prata Porci, especially near its N. extremity, of ancient fragments, marbles, and so on, and of the pavement of the road, which I have marked as doubtfully ancient in the map (cf. Papers, i. 244 and n. 1, supra, 318 and infra, 325). He also noticed traces of an aqueduct of cement, and observed that the hills from this point to Pantano Secco were covered with a stratified calcareous deposit, like that of the Aqua Claudia— that the Aqua Claudia actually passed here {Papers, cit. 245) he did not apparently know. Above the basin on the W. he noticed a small reservoir of selce concrete some 6-6o metres square, probably that N. of point 172 in our map, with a spring near it. Some way to the S. again are the so-called Grotte dello Stingo or Stinco (' Shinbone') so called, no doubt, from the discovery here of some post-classical tomb. The plan of the building— which is the substruction of a villa, with a reservoir in the centre, below the present ground level—is curious, and was noticed by Nibby. Excavations were made herein 1831 by Kestner, Charge d'Affaires of Hanover (the site belonging to Marchese Campana) : see Atti del Camerlengato, iv. 1499, which, however, gives no particulars of any discoveries. To the S. of it, on the N. edge of the railway, on a hill 206 metres above sea-level is a reservoir, now partly ruined, divided into two chambers by square pillars, also noticed by Nibby. This reservoir is marked 'Grotte'in the map. Stevenson {Vat. Lat. 10572,42) tells us that the pavement of the ancient road was found just to the S. in making the new railway line. A little to the E. is a very large villa platform (upon which is built the Casa Boldetti) cut through by the modern railway line, with a supporting wall of opus incertum, the building itself being in opus reticula- tunt: in the hill Stevenson noted ancient pozzolana quarries. From the description of Wells (Alban Hills, 282 sqq.) it would seem that this is the villa which she calls the Villa of Sardanapalus, in which were found in 1761 the statue of the bearded Dionysus, with the name Sardanapalios in Greek characters (now in the Sala della Biga in the Vatican, no. 608), the Caryatides in the Villa Albani (nos. 16, 24, 91, 97, Helbig, Fuhrer, ii. 767- 770), and a very fine draped female figure, with one arm wanting, mentioned by Winckelmann in a letter quoted by Fea, Storia dell' Arte, iii. 253, and Miscellanea, i. 184 (cf. Winckelmann, Werke (Donauoschingen, Y 2 324 THE BRITISH SCHOOL AT ROME.

1825), ii. 97, v. 199, viii. 307. The villa was thought at the time to be the villa of the gens Porcia, from some inscriptions found there, as Winckel- mann says ; but there are none in the Corpus to which his description could refer. It then acquired (I do not know how) the traditional name of the Villa of Lucius Verus. Here it was that Wells copied the fragments of a large inscription, possibly of T. Clodius Eprius Marcellus (Prosopographia, i. p. 415, no. 915), which are given in C.I.L. xiv. 2612. She adds a description of the ruins, and says that five altars of peperino were found here (p. 285), but her account of them is not very clear, and is probably that of her informant. The lead pipe bearing the inscription Aeliae Aste twice repeated, which was found in the railway works in 1891 {C.I.L. xv. 7829), probably belonged to the reservoir marked ' Grotte,' as it was found at the 22nd kilometre (which is close by) ' on the near side' (to Rome) of the Grotte dello Stingo : here were also found (Not. Scavi, 1891, 289) remains of a large, and apparently late, cemetery (inasmuch as burials under tiles prevailed). One inscription was discovered, erected to a priestess of the Bona Dea, with some surprising errors in orthography [F]laviae [A~\thenai [di] Flavins \R~\espaectu\s\ (sic) [m]atri suae b\ene\ merenti feci[t ei\ Marius Alexander et M~\arius Felix ne\potes sacer\doti Bonae De\ae qu~\ae vicshit (sic) a\nnis ] vii d(iebus) ixx (h)or(is) .... \ina\rma- rarius (sic) isc it curae la. Two Arretine vases and a lamp, all with stamps {C.I.L. xv. 5151 a. 5, 5517 b. 6, 6376, 25 (d.) 1 were found, and imperial coins, including a fine one of Alexander Severus, coined in Syria in 230 A.D. (Cohen iv. p. 64, n. 485). Stevenson too notes (id. 144) that Boldetti had found tombs with stamped tiles, no doubt before the railway was made. ' Near Prata Porci' was found the sepulchral inscription C.I.L. xiv. 2712. Further to the E. are the ruins marked Grotte di Colle Pisano in the map, which belong to a villa : at point 195 is a vaulted chamber with two aisles (probably a reservoir) and adjoining it a substruction wall in good opus quadratum of peperino blocks, the courses varying from C41 to C53 metre in height, running S.S.E., with buttresses at the S.S.E. end : the wall is about 40 metres long, and served to support one of the platforms of the villa. To the E. is another reservoir, rectangular, and apparently open. To the S. of this villa the railway cuts through a drain and to the S. again at point 227 is the platform of a villa, with a subterranean chamber in it. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 325

Further W.S.W. at point 229 are the remains of a large tomb—a lofty solid mass of selce concrete, over 5 metres square, the so-called ' Torre della Bella Pisana.' Wells (p. 289) mentions the discovery above it of fine pave- ments and fragments of statues. There are, however, no decided traces of antiquity upon the path which passes by it, and I do not see sufficient grounds for assuming it to be ancient: some pavingstones in the fieldwall at its S. end might also come from the Via di Colonna {infra, 326) ; and a little to the S.E. of the tomb, on the S.W. side of the path, are remains of an ancient building in selce concrete facing S.W., with a drain running N., which probably extended across it. At the junction of this path with another S. of point 261, Stevenson {Vat. Lat. 10572, 22) noted the discovery of an underground water channel in 1892, and also the existence of worn, broken pavingstones. To the W. again, in the Quarto Vagnolo, Stevenson marks ruins: there are the remains of a mediaeval build- ing, possibly on an ancient site, incorporated in a modern house. Here I saw a much weathered cippus in white marble 105 metre high by 070 wide: of the inscription on it I could only read D M S in the first line, in letters 0*07 metre high. Returning to the railway line at the Grotte, we follow the Vicolo di Prata Porci straight to the S.S.E. There are no traces of antiquity along it now, but Grossi-Gondi (p. 156) cites Giorgi's statement that on June 23rd, 1734, he saw the ancient pavement running parallel to this road (Stradone di S. Marco). As given by Lanciani {Bull. Com. 1884, 173,) it does not really justify any statement of the kind—Giorgi simply says there that it was found below Frascati in a tenuta of Prince Borghese called S. Croce below Vigna Spinetta {supra, 318). Nibby {Schede, i. 78), however, mentions that in his excursion from Frascati to Prata Porci and Pantano Secco, after noticing the ancient road under the enclosure wall of the Villa Borghese {infra, 330) he and his companion took the road to the left before reaching the Montano (di Borghese, i.e. the Barco) and that this in some places seemed to him ancient. This must refer to the road to Colonna, and cannot be the same as the road mentioned ib. 104 (cf. Analisi, III. 9) which, he says, left the Via Labicana at Torre Nuova (really rather further on—supra, 322) joined the modern road to Monte Porzio beyond the Cappellette {infra, 375) and ascended to Tusculum on the side of Camaldoli. 326 . THE BRITISH SCHOOL AT ROME.

I am not indeed quite sure about the existence of a road coming up from the direction of Pantano Secco or Prata Porci to the 'E. of the Cap- pellette: Grossi-Gondi does not mark it, nor does Rosa in his large map of the Campagna. The latter has a different theory again, marking an ancient road coming from the bifurcation a little S. of the Cisternole, pass- ing E. of the villa on Colle Spinetta, W. of the Barco di Borghese, and then curving round so as to pass between Mondragone and Le Cappellette {infra, 369). In the Quarto S. Croce, in the Vigna Graziosi, a lead pipe bearing the inscription Ti(berius) Cl{audius) For ( ) fedit) was found about i860 according to one account, or in 1879 according to another (unless two copies of it were discovered : cf. C.I.L. xv. 7859). In a vineyard in the same Quarto were found the unimportant sepulchral inscriptions C.I.L. xiv. 2694, 2731. At the end of the straight line the Vicolo di Prata Porci falls into the road which leads to Colonna on the S. side of the latter. Grossi-Gondi (p. 155, n. 5) indicates a tomb, the so-called Torretta, which I have omitted to mark in my map: it is just under the ' to' of ' Qto,' and Stevenson also noticed it. I have since examined it, but consider it to be a fragment of a larger building, the greater part of which is hidden under the slope of the hill. What is visible is a part of the substructions in opus reticulatum, with a vaulted roof: above it is a ruined modern house and a little further up another. That the road continued past it to the S. is a justifiable assumption, though there are no actual traces as far as the modern road to Monte Porzio. For its further course up to the amphi- theatre of Tusculum cf. infra, 367 sqq. There are no decided traces of antiquity in the road to Colonna E. of this point, and I had {Papers, i. 253, n. 2) expressed some doubt as to the actual line of the modern road representing the ancient one, though I still con- sider that such a line of communication must have existed. Nibby (Ana/isi, ii. 356) is only speaking, however, of the road above Monte Porzio, not of that below it: for his map and his description show clearly that in his time there was no road passing below Monte Porzio to La Colonna (infra, 383) and that he went by the upper road, of the antiquity of which from this point to Frascati there is no doubt; and it will be better before examining the district further E., to return by it and dispose of the villas immediately above Frascati, and of Tusculum itself. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 327

To the N. of the road is the huge villa called II Barco di Borghese (Barco is simply a corruption of Parco—cf. Papers, iii. 122). It is an enormous artificial terrace (there was probably on the E. half a natural elevation which served as its nucleus, but the whole of it is now surrounded by substruction walls), upon which there is now a garden, there being no remains of the villa itself preserved. The W. portion of the platform is supported by a series of substructions of a very remarkable kind : a careful plan of them has been made for me by Mr. F. G. Newton, and will be found reproduced on PI. XXVII. At the N.E. angle there is a projecting mass of concrete foundation which seems to mark the extremity of the villa: the outermost wall marked on the plan, going westward from this point, is not preserved above ground level: the ancient facing of the inner wall, both on the N. and E. sides, exists for a foot or two only, and above it is modern concrete. It is made of irregular pieces of selce (cf. supra, 315, and PL XXXV. fig. 2), but towards the middle of the N. side opus reticulatum of the same material begins. Where the building projects further to the N. there is a cryptoporticus on the outside (cf. Grossi-Gondi, op. cit. PL VI.), while within the platform are the substructions of which we have spoken, which are now accessible from the chamber B and are entirely without light or ventilation. The first range of chambers which we reach runs S.: they are large and lofty, and are faced with rough pieces of selce, the vaulted roofs being made of concrete set on planks: the apertures leading from one to the other have sometimes pointed, sometimes rounded, tops. There are two low internal cross-walls in the last but one, the first of which is faced with opus reticulatum. The isolated chamber to the S., marked A, is accessible only from the garden above. To the W. of these is another group, which at first has more sub- divisions, though the further chambers are even larger than those of the first group ; but another group of small chambers lies to the S. The great thickness of the dividing walls is remarkable, and probably there was a considerable weight above to be supported. The concrete was here faced with brickwork of not very good quality ; but many of the bricks have been hammered away in mediaeval or modern times for building material: some of the walls were faced with opus signinum, so that a few (though a comparatively small minority) of these chambers must have served as reservoirs. To the S. of these small chambers a long corridor runs 328 THE BRITISH SCHOOL AT ROME.

S. for over half the length of the substructions : at C is a wall in opus reticulatum, which does not exist to the full height of the corridor, and here is a branch passage running W., which leads into a corridor parallel to the main one, the entrance into this last being by a pointed aperture formed in the concrete. From this shorter corridor (which may be regarded in one sense as equal in length to the main one, though it is much broken up and subdivided) the outer air may be regained on the W. side of the villa by a break in the back wall of one of the long chambers running E. and W. The main- corridor may be followed S. for a considerable way : we find a single chamber on the E. of it, and a group of small ones on the W. (The inaccessible rectangular spaces between the groups of chambers must have been filled with earth.) Just beyond the entrance of the former there is a vaulted passage inside it, the walls of which reduce its width so considerably that it is now blocked with earth for a distance, as far as can be estimated, of about four metres. If one could pass through this obstacle, one would reach (in some cases only by later breaks in the wall) a number of long chambers going E. and W., and to the S. of these another group of small chambers (in that marked F there is a double vault, with a hollow interval of 1*50 metre between the two), beyond which again, though inaccessible from them, are more long, narrow chambers. To the W. of these last the plan shows an open space, now partly occupied by modern buildings and partly left open ; but it seems to me that there are sufficient indications to lead us to suppose that there was a corridor here, and more chambers to the W. of it: it is improbable that there was a courtyard here in ancient times, owing to considerations of level. The narrow space to the N. of this yard shows similarly clear traces of a series of parallel chambers with vaulted roofs. Returning to the long internal corridor, just before the point where it is blocked, we notice that a group of small chambers lies to the W. of it: traversing these, we find ourselves in the line of what is in reality a parallel corridor to it, though much subdivided, and actually blocked in two places. Following it northward, we pass through a series of small rooms, from which others branch off: these last present curious thickenings of walls and irregularities in plan. Some of them have served as cisterns. In two of them, marked D, is brickwork, with bricks no less than O'4O to 0-50 metre long, and 0-07 metre thick, a thickness, I think, unknown elsewhere in this part of Italy, though it is met with, e.g. at Urbs Salvia (Urbisaglia) in the Marches. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 329

We have not noticed the exterior of the W. side of the platform, which is mostly faced with opus reticulatum of selce, with quoins of the same material. At about the centre of this at E we find almost the only specimen of brickwork that may be seen on the outside: the bricks averageo'O27 metre thick,the mortar courses 0013 thick. Just to the S. of this we may notice two arches, half-filled at a later period with masses of concrete, faced with small pieces of selce with brick quoins, no doubt in order to give further support. Further S. Nibby marks in his plan a half column of opus reticulatum, half-buried, which I did not notice. On the S. side little is preserved, though a fine specimen of opus reticulatum may be seen close to the modern road. The wall at a divergent orientation, shown only in outline, belongs to a modern house, and is probably of modern origin also; but there is a small fragment of ancient walling at a considerable distance further on, not very far from the S.E. angle. Nibby {Schede, i. 105) gives a rough and not accurate sketch plan of the villa; Grossi-Gondi (op. cit. PI. VI., cf. p. 158) gives two views. Stevenson had believed {Cimitero di Zotico, 98) that it belonged to the Quintilii, but the lead pipes on which he based his assumption were really found at Mondragone *(?'«/>-#, 370). Grossi-Gondi, who has examined the ruins carefully, found in the cornices forming the imposts of the vaults five examples of C.I.L. xv. 310, eleven of ib. 1273, and three of ib. 1289 (all of the time of Vespasian). Giorgi noted as found here in 1734 ib. 904 f. 25 (Trajan) 933 a. 4 (middle of first century) 1365. 7 (134 A.D.) : with this would agree the character of the facing—both the opus reticulatum with stone quoins and the brickwork. See Bull. Com. 1898, 333 ; Wells, op. cit. 238. We may therefore with some confidence assign the building to the end of the first century A.D. Nibby {Schede, i. Jj) mentions an excavation here not long before 1822 in theoliveyard to the W. and speaks of the existence of fragments of marble there. To the E. of the Barco is a reservoir which may have supplied either it or the building at the Torretta. It lies on the N. side of the high-road just before it takes a bend to the S.S.E. towards Le Cappellette. It thad three chambers, of which two are now preserved, with nine arches in each of the dividing walls. Grossi-Gondi alludes to it (p. 156, n. 1) as in the Vigna Carocci, but marks it in the wrong place, to the E. of La Torretta. Just to the S. of the real site of this reservoir, in the elbow of the road, E. of the 330 THE BRITISH SCHOOL AT ROME.

Villa Vecchia, he marks the remains of another villa, of which scanty traces exist. Between the Barco and Frascati, near the entrance to the Villa Mondragone, as Grossi-Gondi tells us (p. 144), pavingstones were found on the line of the modern road in 1899-1900; and Nibby tells us that he saw the pavement of the ancient road under the enclosure wall of the Villa Borghese on the left on the steep ascent to Frascati (the former Via Saponara, now Via di Monte Porzio), as Grossi-Gondi and I have marked it (Analisi, ii. 356; iii. 342 ; Scliede, \. 61, 78 : cf. Westphal, Romische Kampagne, 80).

XVII.—THE ASCENT FROM FRASCATI TO TUSCULUM AND TUSCULUM ITSELF. The road which we have followed from II Barco leads us to the level of the cathedral of Frascati. From this level, according to Grossi-Gondi's researches, three other roads branch off {supra, 301), one leading to the amphitheatre of Tusculum and passing E. of the Villa Falconieri, another going to the same point, but passing W. of the Cappuccini, and a third, with which we have already dealt {supra, 252), leading S. to the Via Latina at La Pedica near the thirteenth mile. The first of these I have conject- urally marked as starting from near the Camposanto of Frascati {supra, 316) owing to a misunderstanding, as Grossi-Gondi points out (p. 156 n. 3), of information which he gave me. Grossi-Gondi, on the other hand, makes it follow a different course : it starts from Frascati, then passes W. of the Villa Borghese, and runs up the N.E. side of the valley which bears the name of Valle di Ciceronel (infra, 339) and leads up to the amphitheatre: along the bottom of it runs the boundary between the communes of Frascati and Monte Porzio. He describes various remains of pavement which have come to light in recent years, from which its width can be determined at about 2 metres. Above the modern path which runs from Frascati to Camaldoli, indeed, it can be traced for about a quarter of a mile. To the S.W. of this road rises the Villa Falconieri, previously known as Rufina, from its builder Alessandro Rufini, Bishop of Melfi : it was also called Villa della Maddalena from a chapel2 destroyed in building the 1 For a find of antiquities here in 1850-2 see Seghetti, Tuscolo e Frascati, lot, n. 1. 2 In this chapel was copied the inscription C.l.L. xiv. 2609. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 331 villa, which was the first of the modern villas of Frascati (1546-50) (Lanciani, Sioria degli Scavi, iii. 45). It has recently become the property of the German Emperor. It rests upon an ancient villa, of which some remains are still visible—especially the massive substructions on the N.W. and N.E., in opus reticulatum (carefully described by Stevenson {Vat. Lat. 10572, 27" sqq.), while several walls of this material, and plain mosaic pave- ments with yellowish tesserae, one with a narrow black band round the edge, have been recently found in the garden in front of the villa. On an upper terrace to the S. is a beautiful pool, the so-called Ninfa, of which Lanciani gives an attractive photograph in his Wanderings (after p. 286) : the reservoir at the W. corner is not ancient, for an opus reticulatum wall projects close to it, showing that the ancient villa should have been marked in my map as extending further S.W., as far as the modern entrance gate. An anonymous letter of the 8th of August 1753 preserved in the Chronicon Sublacense of the Library of the Episcopal Seminary at Frascati, mentions the discovery of, apparently, a herring-bone pavement on the upper level, where the fountain now is {Bull. Com. 1884, 204). That Rufini found antiquities in building his villa is probable, inasmuch as Aldovrandi (p. 181) saw in his house in Rome a female head found at Frascati. Here Fabretti and Lesley late in the seventeenth and early in the eighteenth century saw the inscription C.I.L. xiv. 2604. It was erected to M. Cusinius M. F. Vel{ina tribii) aed{ilis) pl{ebis) aerario praeifectus) priaetor) by his father, his mother Fictoria, and his sister. Dessau marks that the absence of cognomina proves the inscription to be earlier than Nero; and Klebs {Prosopographia, i. p. 488, no. 1329) assigns it to the years 27-23 B.C., cf. Tac. Ann. xiii. 29. If the inscription was found here (which of course we do not know) then there is some ground for supposing that the villa belonged to this person. In the first half of the eighteenth century C.I.L. xiv. 2633 was copied here: it was a dedication to Julius Severinus patronus cultorum Dianesium by three other persons, two of them slaves, which seems to be of late date. The inscription C.IL. xiv. 2656, cut upon a cippus of sperone was also seen here in the seventeenth and eighteenth centuries, and was copied by Stevenson, who saw another similar cippus without inscription : he noted that it was not fractured on the left, as represented by Giorgi. The lettering is not very distinct, but the reading given is correct, except that xvii in the last line should probably be xviii. In a niche below the Ninfa is the headless torso 332 THE BRITISH SCHOOL AT ROME. of a male statue wearing a toga, and a Doric cornice with a bucranium between two triglyphs. In the hall of the Villa, too, are several Corinthian capitals of white marble, and in the lower garden is a statue of a man bearing fruit in his bosom. In the fountain was recently found a slab of white marble C09 metre thick with the following inscription, which was somewhat difficult to decipher: it is obviously sepulchral and of little importance. 1 o-io \NTI 'IVSI v 0-05 062 metre CLEMEN 0-045 ^TSIBI C04 A XXII 0-037

0.38 The various remains of villas, etc. further E. will be described in connexion with the continuation of the road from Prata Porci {supra, 326) and in the reverse direction, i.e. starting from the amphitheatre {infra, 367). We now return to Frascati and take the road which leaves it on the S.S.E., and which is, I think, undoubtedly ancient, though Nibby says {Schede, i. 64) that it is quite modern. On the W. side of this road, before we actually leave the modern Frascati, we find a tomb which tradition attributes to Lucullus (cf. p. 302). A plan of it is given by Angelini and Fea, Via Latina, PI. VII ; and another less accurate and differing, in that the exterior is made square and not round, by Canina, Tuscolo, tav. xxvi. (with a view ibid, xxvii) = Edifisi, VI. tav. lxxxiii. It consists of a circular mass of concrete with a chamber in the form of a Latin cross in the centre, faced with opus reticulatum with brick quoins. Mattei {Tuscolo, 61, 62), who gives a rough wood cut of it, says that it was despoiled of its decorations about 1598, during the construction of the new cathedral, though the best of the objects found were removed to Rome. Close by it in his own time some burials covered by tiles were found. Opposite to it is the Villa Lancellotti, formerly S. Croce, S. Angelo, and Piccolomini, which has no archaeological records (Lanciani, Storia degli Scavi, iii. 56). The inscription C.I.L. xiv. 2708 (of no interest) preserved there was very likely found in Rome and transported thither. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 333

The path, between high walls, retains some traces of ancient pave- ment : it passes to the N.E. of the splendid Villa Aldobrandini, which was begun in 1602 by Cardinal Pietro Aldobrandini, nephew of Clement VIII. (Lanciani, Storm degli Scavi, iii. 55), and which in the 18th century was called the Villa Pamphili.1 This does not itself occupy an ancient site ; but, immediately to the S.W. of it, between it and the gate towards the Villa Montalto, are foundations of selce concrete which form angles with the avenue, and must be corners of cisterns or of rooms in the substructions of an ancient villa which was orientated differently from the present one. Near the gate just mentioned there was found in 1845 a sepulchral cippus, with two urn-shaped cavities for ashes, bearing the inscription C.I.L. xiv. 2610 (Grossi-Gondi, p. 191, n. 1). It is an inscription in honour of Rubellia Bassa, daughter of C. Rubellius Blandus and of Julia, the granddaughter of Tiberius (Prosopograpkia, iii. p. 136, no. 86), set up by her grandson Sergius Octavius Laenas Pontianus, consul in 131 A.D. It is possible that we have here an indication of the ownership of the ancient villa. For another property owned by the same man at Prata Porci cf. Papers, i. 244. There were also found on the same occasion, Kaibel, I.G. xiv. 1003 (a marble base with a metrical dedication of a statue holding a cup to Hercules), a Latin inscription of Certa (which does not seem to have found its way into the Corpus), a trapezophorus (cf. Grossi-Gondi, 73, n. 1), an altar with trees and birds (still in the modern nymphaeum), a fragment of a draped statue, and a piece of a coffered ceiling (Atit del Camerlengato, iv. 3453, cited by Tomassetti, Via Latina, 180 n): ib. 3475 contains papers relating to the continuation of the excavations, but there is no information as to their results.2 Pavingstones are to be seen in the substruction wall of the garden near the villa, which may have come from the road up to the amphi- theatre, while others, Stevenson thinks (Vat. Lat. 10572, 134), which are to be seen to the S. along the avenues which lead along the W. side of the hill, may have been brought from elsewhere, or may belong to another deverticulum, which would have fallen into the branch from the Via Latina to Tusculum {supra, 257). 1 Cf. Eschinardi, Descrizione di Roma (ed. Venuti, 1750), p. 265. 2 Tomassetti cites ib. 2846 as his authority for the statement that with the inscriptions were found five marble heads in 1834, but incorrectly {infra, 344, n. 3). 334 THE BRITISH SCHOOL AT ROME.

Whether many of the antiquities now or formerly preserved at the Villa were found at Frascati or not is as doubtful as in the case of other villas at Tusculum. They include the famous Demosthenes of the Vatican (Braccio Nuovo, 62), a bust of Vespasian (possibly not genuine—Bernoulli, Rom. Ikon. ii. 2, p. 23, no. 10), a statue formerly supposed to represent Domitian, but really a Hermes of the neo-Attic school of the first century B.C. now at Munich (Glyptothek, no. 300), having been bought in Rome in 1811 from one Pescetelli, an inferior male sepulchral statue, in white marble, still in the modern nymphaeum, as Tomassetti calls it {i.e. the large hemicycle just to the S.E. of the Villa itself), and three or four sarcophagi, which were in Rome in the sixteenth century, and removed to the Villa before 1656. (Robert, Sark. Rel. ii. 34 ; iii. 79, 155, 224.) A bust of Volusianus is mentioned by Bottari, Mus. Cap. ii. p. 67, cf. Bernoulli, op. cit. ii. 3, 161 fin. An inventory of the collection of sculptures, as it was in 1709, is published in Documenti ineditiper servire alia storia dei Musei d'Italia, iii. 181 sqq. The Villa is supplied with water from springs below Monte Fiore {infra, 413), which is brought by an aqueduct constructed for the purpose by Cardinal Pietro Aldobrandini. Illustrations of the fountains and gardens of this and other villas at Frascati at the end of the seventeenth century will be found in Falda's Fontane delle Ville di Frascati; see also- H. Inigo Triggs, The Art of Garden Design in Italy, pp. 117 sqq. Pis. CII. sqq. I shall not deal with the fascinating subject of the Renaissance villas of Frascati ; though this should be done with reference also to the history of the collections of antiquities which they contained. The unimportant sepulchral inscriptions C.I.L. xiv. 2684, 2685, 2730, 2734 were found in the grounds of the Villa, probably all of them in 1875 (cf. Wells, op. cit. 193 sqq.) along the deverticulum mentioned, supra, 258, and several others built into the so-called Casino delle Fornaci, i.e. the house between the Villa and point 429 (where are many fragments of marble sculptures arid decorative terracottas) were probably discovered there also (ib. 2643, 2644, 2751, 2720, 2729, 2735, 2739 and the Greek inscription given by De Rossi, Bull. Crist. 1881, 131). As to ibid. 2743 the case is doubtful. Tomassetti (Via Latina, 181. note) notes the discovery, a little before he wrote, of a marble statue and a fine pavement, and speaks of other antiquities visible in the Villa. A little way above the Villa Aldobrandini, on the N.E. side of the CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 335 road, is the monastery of the Cappuccini,1 which occupies the site of an ancient villa, described in the extract from Cod. Tusc. cit. given by Lanciani in Bull. Comm. 1884, 202, and Storia degli Scavi, iii. 47: it had a reservoir close to the ancient road, and behind this a substruction wall, in which in 1656 the anonymous author of the description cited saw seven niches, decorated with shells and calcareous incrustations (the so-called ' tartari') and mosaics. This substruction wall supports the terrace of the Villa Rufinella, which thus occupies probably the site of the main building of the villa, which I have omitted to mark in my map. This villa was only separated from the Rufina (Falconieri) in 1578; but the subsequent changes of proprietors are given quite differently by Lanciani (Storia, loc. cit.) and Grossi-Gondi (p. 149, n. 3). In 1804 it belonged to Lucien Bonaparte, Prince of Canino, in 1820 it passed to the Duchess of Chablais, then to the royal house of Sardinia, and finally to Prince Lancellotti. The ancient villa which is associated with the name Rufinella is one which lies within its grounds, but some 500 metres to the S.S.E. of it, 538 m. above sea-level, upon the boundary line between the communes of Frascati and Monte Porzio, and mainly in the territory of the latter. This is the site of the discoveries of 1741-6 (during the period 1740-73, when it was the property of the Jesuits) described in full by Lanciani, Bull. cit. 174 sqq., and Grossi-Gondi, 148 sqq., both of whom give full references to previous authors. At present there is very little visible above ground : the two platforms of the villa, which seems to have faced S.W., are clearly distinguishable (though, the walls having been razed to the ground as they were discovered, they are now occupied only by flat meadows), and scanty remains of the substruction walls of the lower in opus reticulatum of selce with quoins of the same material may be seen on the S.W. and N.W. sides; while on the upper terrace, between the road leading up to the amphi- theatre and the boundary wall of the Aldobrandini property, are three reservoirs, two with one chamber each, and a third with at least three. Unfortunately the plans and notes made by Boscovich with a view to a thorough publication of the results have disappeared, and the accounts at our disposal are somewhat meagre. The devastation of the site is

1 On the threshold of the entrance door to the monastery Lanciani copied a fragment of an inscription (C.I.L. xiv. 2764a). 336 THE BRITISH SCHOOL AT ROME. accounted for by the fact that the walls were demolished to provide material for the new wing of the villa, which was being erected by Vanvi- telli. The villa had four courts, according to one authority (Giornale dei Letterati 174.6, 115 sqq., reproduced by Fea, Miscellanea, ii. 130 sqq., and Canina Tuscolo, p. 87, n. 5), two according to Zuzzeri (Di mi antica villa scoperta nel dorso del Tuscolo) and Lanciani accepts the latter number as more in conformity with the usual type of a large country house. The baths were clearly recognizable, and were on the N. side, where the principal rooms seem to have been situated, for the sake of coolness in summer.1 There was a peristyle on three sides of the whole villa, with a crypto- porticus ; and there were several cisterns (including those we have just mentioned), near which numerous water-pipes were discovered, though only two inscriptions have come down to us, T. Cispius Decor (C./.L. xv. 7857) and L .... Rat. Sc.... {ib. 7872), both given by Zuzzeri. To these I feel no doubt that we should add ib. 7839 N. IuniFaustiand 7874 L. Aennius Fortu- natus, both given by Muratori from copies sent him by Ramagini, who as we know {supra, 224) was copying inscriptions at and near Tusculum in 1741, and who saw them ' Tusculi in aquaeductu! The copy of the latter is no doubt corrupt, and there is some doubt as to what the gentilicimn should be. I think we may very likely also add the terracotta friezes in the Museo Kircheriano, one of which is published by Winckelmann, Mon. Ined. 161, while both are given by Canina, Tuscolo, tav. iii. iv. with the mourning Penelope and the recognition of Ulysses. The brickstamps found here are given by Zuzzeri {pp. cit. 26 sqq.): they are as follows, C./.L. xv. 371. a. 19 (Severus),2 566. b. 2 (about 123 A.D.), 571. 1 (Hadrian), 595. a. 8, b. 32 (Hadrian), 861. 15 (142 A.D.), 966. d. 7 (first century A.D.), 2042. b (123 A.D.), 2233 (first century A.D.), 2277 (first century A.D.).3 The fine head of Socrates now in the Villa Albani (no. 1040, Helbig, Filhrer, ii.2 834; Bernoulli, Gr. /kon. i. 187, no. 8) was discovered in the Villa of Cicero ' situated on the summit of the ancient city of Tusculum ' in 1735, according to Pier Leone Ghezzi, Cod. Ottob. Vat. 3108, 167 = Lanciani, Bull. Com. 1882, 224, lxiii. = Schreiber, Fundberichte des P. L. Ghezzi, lxxiv. in Sachs. Berichte, 1892, 141). It first passed into the hands 1 Winckelmann in Fea Storia delVArte, iii. 83, 211, ( = mem. 10 Fea, Miscellanea, i. p. 203) gives a long description of the hypocausts. 2 This is doubtful (cf. supra, 301). 3 For this stamp which bears the legend M. Tuli see supra, 234. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 337 of Cardinal Alessandro Falconieri, whose heirs sold it to Mgr. Niccolini, who in turn gave it to Cardinal Alessandro Albani. I have already pointed out {Papers, iv. 115) that Ghezzi gives a different account of the locality of the discovery of the other objects which he associates with -this head ; but I now think that the fact that he says expressly ' I was present at the discovery of it' must have more weight than I had given to it. The fine mosaic in the Vatican (Sala in forma di Croce Greca : Helbig, Ftihrer, i.2 328) was found in 1741 in this villa, cf. Visconti Museo Pio Clem. VI. tav. xlvii., Canina, Tuscolo, tav. xliv., Grossi-Gondi, PI. VII. for illustrations of it. Other small fragments of mosaic found in the same year, but not certainly belonging to the same pavement, were in the Museo Kircheriano, and are now in the Museo delle Terme (Canina, tav. xiv. Helbig, ii.2 1153, 1157): cf. Ficoroni, Mem. 74 in Fea, Miscellanea, i. p. 153. Ficoroni, in Roma Antica, 1741, 275, mentions the discovery here in the previous year of two bas-reliefs with two Bacchantes, one holding a cantharus and one a thyrsus, which after they had been restored and a cornice 7 palms high added were conveyed to France by the Duke of S. Agnan, ambassador of the French King to the Holy See. Fea, Miscellanea, i. p. 153, mem. 69, leaves out the word ' bas-reliefs.' A horologium solare was also found and is described and illustrated by Zuzzeri, op. cit. 63 sqq.: it too was placed in the Museo Kircheriano (De Ruggiero, Guida del Museo Kircheriano (1879) p. 60, 224-226). Dessau is almost certainly right in attributing to these excavations C.I.L. xiv. 2635— Plutiae A(uli) f(iliae) Olympiadi sodali iuvenum l(ocus) d(atus) diecurionuni) d(ecteto)—and 2640 (sodales lusus iuvenalis). He notes, rightly, that it is only at Tusculum (id. 2631) that we find women as sodales iuvenum, and notes further that copies of both inscriptions were sent to Maffei by Contucci just about the period of the excavations (at least before 1749) and that the second is still in the Museo Kircheriano. There seem to be faint records of earlier excavations, for a large marble cippus bearing the inscription C.I.L. xiv. 2607 is described as having been discovered in 1727 or 1728 ' on the summit of Tusculum,' or more precisely ' near the road which leads over the top of the mountain from the Vigna Pamphili (i.e. the Villa Aldobrandini, supra, 333) to Mondragone': very probably at this villa itself, but in any case somewhere between it and the amphitheatre. 338 THE BRITISH SCHOOL AT ROME.

The cippus was set up in honour of M. Gavins T. f. Veliina tribii) Appalius Maximus, clarissinius vir, quaestor Augusti, legatus provinciae Narbonensis, sodalis Hadrianalis by a freedman and procurator of his ; and, if it was actually found in the remains of the villa, it may give us the name of one of its owners in the latter part of the second century—a date which would agree with that of the brickstamps ; but the use on the S.W. side of the lower terrace of selce concrete faced with opus reticulatum with selce quoins makes it probable that the villa had existed a good deal earlier, though there is no ground for attributing it either to Cicero or to Tiberius [supra, 232). Lanciani {Bull. com. cit.) is in error in attributing to this site the excavations in 1831 and 1834, which belong to the building which Canina calls the Villa of Tiberius {infra, 344). Of subsequent excavations in this part of the Villa Rufinella we have no record, except that we are told that some brickstamps were found in digging to plant chestnuts in the Villa Lancelotti (Rufinella) namely, C.I.L. xv. 454 c (123 A.D.) and several others, mostly rectangular with one line of text, belonging, probably, tc the first century A.D. (Tomassetti, Bull. Com. 1890, ill). E. of point 538, above the winding avenue, is a substruction wall of selce concrete facing S.W., which I am inclined to attribute to the Renais- sance period ; and further to the E. again are some rough ruins, which may be mediaeval. I have not marked either in the map. As we continue to ascend E.S.E., the pavement of the ancient road begins to be preserved a little to the right of the modern path, not long before we reach the W. side of the amphitheatre of Tusculum. Here our path is joined by three other ancient roads, one ascending on the E. side of the Villa Falconieri {supra, 330), another, the prolongation of the road from Prata Porci {supra, 326), ascending past the Villa Mondragone {infra, 368), and a third ascending from the Via Latina at La Pedica {supra, 260). See the block plan of Tusculum (PL XXVIII) which has been prepared from Canina's plan {Edifisi, VI. tav. 85 = Tuscolo tav. vi.) by Mr. F. G. Newton, with various corrections made on the spot, and with the omission of much of the detail, simply in order to serve as a guide to the text. The best plan of Tusculum, as far as it goes, is that in Rossini, Contorni di Rovia (1826), tav. 71. Angelini and Fea, Via Latina, tav. viii, also give a good plan dating from 1828. Both of these were of course made before the CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 339 completion of the excavations, but have the advantage that they contain none of the arbitrary restorations which figure so largely in Canina's work. A plan of Tusculum, on the scale of 1 : 1000, with elaborate recon- structions, was exhibited by M. Gamier at the exhibition of drawings of the students of the French Academy at the Villa Medici in 1903, but has not been published. The amphitheatre (no. 1 on our plan) was absurdly called by the earlier antiquaries the Accademia or Scuola di Cicerone, and still bears that name on the Staff Map. Plans, sections, and views of it are given by Uggeri, op. cit. tav. viii. and Vedute, no. 14, Angelini and Fea, op. cit. tav. ix., Canina, Tuscolo, tav. xxii. and x.x\i[.=Edifizi, VI, tav. xc. The diameters of the arena are about 47 and 29 metres, the total diameters about 80 and 53 metres. The first excavations were, as Uggeri says (text, p. 30), made by himself in 1820, while Canina, op. cit. p. 132 n. 1, tells us that he was responsible for the plan, which was also reproduced by Angelini and Fea. The plan in Canina's own work is a good deal more complete. The building has, I should say, never been entirely cleared,1 and a good deal has been filled in since the excavations. A view of the interior as it stands, looking towards the S.E., is given in our PL XXVI, Fig. 2. The site chosen is the head of the valley which descends between the Villa Rufinella and the Villa Mondragone, so that on the E.N.E. it is supported by the slope of the hill, while on the lower side substructions are necessary, and the only entrances on the level into the arena are at the N.W. and S.E. ends,2 while the seats were only accessible from above or from the arena (stairs are preserved on each side of the archway on the S.E. but are not well shown in any of the plans) and not, as is more usual, from passages concentric with the arena running under the seats. In the eastern half of the circumference such a passage does exist, but it is com- pletely closed, and serves merely as a support; while in the other half it is replaced by semicircular chambers which serve in the same way as supports, and were almost entirely closed. They are now filled up and inaccessible. The use of such spaces may be seen on a much larger scale in the

1 Wells (p. 193) speaks of excavations by Prince Aldobrandini in 1867. 2 The external wall is clearly visible on the S.E. showing that it was not built against the hill here, but only on the N.E. and S. W. sides. This is not clear in Canina's plan, Z 2 340 THE BRITISH SCHOOL AT ROME.

theatre of Augusta Raurica, where they are entirely closed (see Mitteilungen der hist. ant. Ges. zu Basel, N.F. II. Das Romische Theater zu Augusta Raurica by Th. Burckhardt-Biedermann, Bale, 1882) and in the amphi- theatre at Trier (both are given by Durm, Baukunst der Etrusker und Ro'mer, p. 218, fig. 222). In regard to the date of the building, Grossi- Gondi (p. 173) gives us an important piece of information : the parts that are visible are in the main, as our photograph will show, of opus quadratum in the lower part, while the arched substructions and the walls supporting the seats are of opus reticulatum of tufa and selce with bands of not very good brickwork; x but in one of the arches of the vaulting of one of the covered passages lie has seen three examples of the brickstamp C.I.L. xv. 622, which dates from the middle of the second century A.D.; and ib. 1370. 2 (belonging to about the same date) was found here by Ficker. Rossini indeed shows {Contorni di Roma, tav. 73) an ancient road running beneath it, across its minor axis, which would be, I presume, a connecting line between the two roads which, after its construction, had to pass round the outside of it—that ascending from the Rufinella and that descending to Mondragone. Some fragmentary inscriptions on the seats are given in C.I.L. xiv. 2606 : one of them, with the letters NIANO, is still lying in a chamber at the S.E. end. In 1885, in clearing away some earth that had fallen, a headless male statue in travertine draped with a toga was found between the amphitheatre and the road leading up to the city: it had no artistic merit. {Not. Scavi, 1885, 477.) In the entrance passage at the N.W. end we may see clearly from the change of construction (the earlier work being faced partly with opus quadratum of tufa and partly with rough thin pieces of selce, while the later facing is of opus reticulatum) that the first ten metres of it have been built later than the rest: this portion falls outside the line of the external walls and must have been added to provide a passage round the outside of the amphitheatre. Canina sees here the remains of a monumental entrance, which to me is doubtful. A plan and elevation of an amphi- theatre (without legend) by Antonio da Sangallo the younger (Uffizi, 1089) are attributed to this amphitheatre by Ferri {Catalogo dei Disegni di Architettura, p. 79) partly owing to the general similarity of paper

1 The bricks are lightish red, rather uneven, and from 17 to 27 cm. long ; test measurements gave an average thickness of 0*031 m. on the N.W. and 0^036 on the S.E., the mortar courses in both cases averaging o-oi6. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 341 and drawing with ib. 1064. Upon this we find a drawing of ' one of the seats of the Coliseum of Tusculum,' of which I give a sketch, kindly sent me with other information by Prof. Hiilsen : the legend states that it is measured in ' dita ' of about O'O2 metre. We may notice here the small projection at the end of the step, which recurs in the schematic drawing of the seats on ib. 1089, and is another ground for Ferri's identification. The ' archo quadro,' on the other hand on ib. 1064, has nothing to do with Tusculum, but is the so-called Janus Quadrifrons near San Giorgio in Velabro in Rome. Volpi, Vetus Latium, viii. tab. ii. fig. vii, shows some of the steps of the amphitheatre. The fact that the arena of the amphitheatre was accessible on the level from the S.E. has considerable importance for us : for it clearly indi- cates that the neck of land connecting the ridge which we have so far been following with that which was occupied by the town was a very narrow one, and that there was originally a depression here, rendering the position more easily defensible. I think we may take it as almost certain that the site of the amphitheatre marks the W. extremity of the primitive city. In 1735 some sepulchral inscriptions of no importance were found ' on the hill of Tusculum towards Santa Maria della Molara,' or ' on the W. (sic) slope of the ancient Tusculum,' or ' below the so-called Scuola di Cicerone [the amphitheatre] on the right' (C.I.L. xiv. 2678, 2697, 2727, 2749). With the sepulchral inscriptions copied at the Villa Rufinella I shall not deal; but I may make a few additions from Nibby's MS. notes. He gives (Schede, i. 64) a copy of C.I.L. xiv. 2683, which he states to have so far been of sperone, and which he saw (with 2579) on the road between the amphitheatre and the theatre on October 13th, 1823 (?): it is as follows :

Q • BAEBIO Q . F FVFIO COE NVI VIR RVTILIVS

In 1871 the sepulchral inscription C.I.L. xiv. 2756 was found 'between Frascati and Monte Porzio in a place called La Molara.' I should imagine that this referred, not to the villages themselves, but to their 342 THE BRITISH SCHOOL AT ROME. communal boundaries: for I do not know of any place called La Molara on the N. side of the ridge of Tusculum. It will be seen from the map that this boundary passes along the ridge of Tusculum from point 55° as far as the amphitheatre: so that the locality would be much the same as that of the discoveries of 1735. Just beyond the amphitheatre are the remains of a very large building (2 in our plan), of which we only see the substructions of the S.W. side, and the great platform which it occupied. We again have a plan drawn by Canina in the work of Uggeri {op. cit. tav. ix., Vedute, no. 15 ; the plan is repeated by Angelini and Fea, op. cit. tav. x.) ; this only shows the portion of the substruction which is actually in existence, while Rossini loc. cit. has an independent plan. Canina, Tuscolo, tav. xviii.-xx. ( = Edifizi, VI. xci., xcii—xciii being a restoration of the villa which is contained in this work alone), gives an elaborate reconstruction of the whole as a great villa : it is based he says (p. 128), on the knowledge gained in excavations of 1830 made under his own direction.1 He also gives a plan of the substructions which goes somewhat further than that of Uggeri, and is in some respects more accurate and more detailed, having been made after the excavations : in others it suffers, as usual, from arbitrary reconstruction. The construction of the building is in opus reticulatum with stone quoins and voussoirs and bands of brickwork: the exterior of the N.W. side is shown by Grossi-Gondi (tav. ix.) and in our PI. XXIX, fig. 1 ; while the massive substructions on the S.W., also of opus reticulatum, have arches of tiles—in one case the broken flange of a tile 2 is clearly seen. Below the N.W. portion of the substructions is an underground passage reached by a staircase descending from the northeasternmost of the rooms preserved on this side, which I have explored. Its walls are faced with opus reticulatum, and it has a barrel roof of concrete which has been set on planks: it is 1 -8o metre high and o-6o wide. After it has run 20 metres S.W. by S., a branch goes off S.E. by S. for ten metres (half way along it is an opening 070 wide in the wall) and then turns back N.E. by N. for about 100 metres. The first passage, after another 40 metres, reaches the opening of a passage or drain in opus quad- ratuni 2'oo metres high and C65 metre wide, with a flat roof of blocks of opus quadratum: it is obstructed by earth and cannot be followed far. Its 1 This is reproduced by Grossi-Gondi, op. cit. tav. ii. 2 The facing 'bricks' average about O"i8 m. long, and are really cut triangles from larger bricks or tiles : they average C036 m. thick, the mortar being o'oio. m. thick. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 343 object is not very certain, but it is most probably connected with the drainage of the building. On the upper level of the platform there is now nothing definitely visible but a large concrete core, just where Canina marks a huge peristyle, about 4 metres high,1 and some 11 metres wide by 10 metres long, with another, some 11 by 6 metres, at an interval of 2 metres to the S.W. The cores have clearly been surrounded by large squared blocks of tufa, one of which, 1*07 wide by cra.5 metre high (the length is uncertain) may still be seen on the S.W. of the main core ; and the interval was no doubt filled in with blocks which have been removed for building material. We thus have before us, it seems to me, the podium of a temple facing S.W., somewhat small, perhaps, in proportion to the size of the platform on which it stands. With this would agree the discovery here (at an unknown date in the latter half of the 19th century) of C.I.L. xiv. 26212—a base of sperone with a dedication by the aediles quinquen- nales made from the sums realized from fines (aere multaticio), and also the existence on the site of a large pulvinar of white marble, belonging no doubt to an altar. The site of the building before us seems to me eminently suited for that of one of the chief temples of Tusculum. That it is outside the line of the city walls, as traced by Canina, proves nothing {infra, 350); indeed I think the ancient city would more naturally have terminated W. of this building rather than E. of it, in so far as one can argue from the conformation of the ground. That it is the temple of Castor and Pollux seems improbable {infra, 355), and in that case we may assign it, if we will, to Jupiter (Liv. xxvii. 4; cf. Macrob. i. 12). From Livy we learn that in 210 B.C. the summit of the Temple of Jupiter at Tusculum was struck by lightning, and that almost the entire roof was removed ; while Macrobius testifies to the importance of Jupiter at Tusculum : 'there are some who record that this month (May) came into our calendar from that of Tusculum, in which Maius is still called a god, who is Jupiter, so-called, that is, from his greatness and majesty.'

1 There is a horizontal line in the concrete only a little way above the bottom of the core as now visible, which may mark a difference of date, though if we suppose all above this line to be an addition, we shall get a very low podium. The mortar is gray brown in both cases, and lumps of selce are used ; but that below the line seems to be finer and darker. The line is actually marked by a thin layer of hard cement C005 m. thick. 2 The fragment of an inscription on a marble epistyle [id. 2645) which may still be seen close by (infra, 347) contains too little to be of any use to us. 344 THE BRITISH SCHOOL AT ROME.

The building figured in the sixteenth century as the Villa of Cicero, but was assigned, after the discoveries near the Rufinella {supra, 335), to Gabinius. Various excavations were carried on there in the first sixty years of the last century.1 First come the excavations of 1826, conducted by Biondi, in which a seated statue was found near the villa, along the road (which Canina calls the Via dei Sepolcri) on the N.W. side of it:2 it was headless, but was so like the statues from Veii and Piperno, now in the Vatican (Chiara- monti, 400, 494), that a head of Tiberius was given to it; and the view that this was the Villa of Tiberius mentioned by Josephus gained further colour from the discovery. The arguments are, however, quite insufficient, as we have already seen (supra, 317), and Grossi-Gondi is quite right (pp. 151, 168) in refusing to accept them. The statue was, according to Canina, transported to Aglie (p. 139, tav. xxx.). Biondi states that here was found also a fine statuette of a seated boy, who is trying to hold a young pig, which is attempting to escape, and two herms, one of them double. Biondi's excavations did not touch the villa itself, inasmuch as it lay in the tenuta della Molara, not within the property connected with the Villa Rufinella; and excavations seem to have been begun in the former property in 1829, though rather further E. In Bull. Inst. 1829, 124, we find mention of a fine bust of a youth recently found in the Tenuta della Molara, and in id. 246 we hear of excavations close to the theatre in which two busts, one male, one female, were found. These are not the busts of Sappho and Corbulo found under Canina's direction in that portion of the lower city which lay in the Tenuta delle Molare on one side of the Forum, which he mentions on pp. 79, 147, and figures in tav. xxxvi., for these last were found in 1834.3 They are now in the Villa Borghese. The so-called Corbulo is not mentioned as such' by Bernoulli; and I cannot help thinking that, though it might have added 1 Uggeri, p. 39, suggests that, as there were some altars and statues in the Borghese collection said to have come from the Tusculan villa of Cicero, excavations may have been made there in the time of Paul V. Of course not everything from the Villa Borghese was removed to the Louvre. 2 The locality of discovery is given by Canina (Tuscolo, 139) from Biondi's notes in the Biblioteca Alessandrina or della Sapienza (the library of the University of Rome) Cod. 106 E, f. 4V. These notes include the beginning of a systematic description which unluckily never got beyond the first chapter. Canina, however, made free use of them. In his reports to the Camerlengato (Alti, Tit. iv. fasc. 246) Biondi gives no details as to the exact locality of his discoveries. s From AM del Camerlengato, Tit. iv. fasc. 2846, we learn that in 1838 Canina found 'in the upper part of the Tenuta della Molara, below the city and citadel of Tusculum,' eight heads, four cippi, a sarcophagus with strigil markings, etc. {supra, 333 n. 2). CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 345 to the bulk of his work, it would have added immensely to its utility, had he given a list of the statues and busts with which he deals under their false denominations as well as under their true ones. As it is, if later criticism has led to a change in the naming of a piece of sculpture, it is often almost impossible to find out what has become of it, if one only knows the erroneous name given to it at the time of its discovery. The regular excavations, described in Atti del Camerlengato, Tit. iv. fasc. 1219, seem to have been carried on in this villa itself. Prince Camillo Borghese, who was trying to found a new museum to take the place of the collection which had previously adorned the Villa Borghese in Rome (of which the greater part had been in 1792 transported to the Louvre, where it still is), had asked for leave to search for works of art in various properties which belonged to his family. In January, 1830 {Atti cit. 2192 barely mentions work in 1834), the Prince obtained permission to excavate in the Tenuta della Molara, which is described as precisely above and on one side of the Rufinella: the work, under the direction of Giuseppe Spagna, a silversmith and metal founder of Rome, was begun near the amphitheatre in March, 1830, and before the 25th a piece of plain mosaic had been found, but broken up furtively. On April 1st it was reported that a bust and a head of marble, both of good style, were found, and leave was asked that with the fragments of mosaic they might be removed to Villa Borghese. In May permission was sought to transport to Spagna's studio in Via del Babuino a piece of black and white mosaic with geometric figures and various fragments of marble. These documents do not mention Canina as in charge of the excavations, though it seems to me impossible not to refer to them his words in Ticscolo, p. 128: 'the excavations directed by me about ten years ago (the book was published in 1841) by order of Prince Borghese Aldobrandini, the site on which these remains exist being included in the Tenuta della Molara, which belongs to the Aldobrandini family.' Some years later, in 1859, other excavations were carried on here by Campana.1 Of these we have no particulars except that various brickstamps found here are known from the copies given by De Rossi to Descemet. Some belong to the first, others to the second century A.D. C.I.L. xv. 562. 18 (134 A.D.), 593. 11 (several copies—Hadrian), 809 a. 4

1 Was it in these excavations that the statues of the four seasons in the Villa Aldobrandini (Wells, 185) were found and the frescoes which were in his collection ? 346 THE BRITISH SCHOOL AT ROME.

(first century A.D.),1 912. 22 (four copies—first century A.D.), 1315 a. 2 (first century), 1527. 2 (first century), 1821 (134 A.D.), 2029 (134 A.D), 2281 (Hadrian ?). In 1866 still further work seems to have been done in the building: for it was in this year that the brickstamps C.I.L. xv. 7 a. 1 Reipublicae Tuscidanor{iim), 1122. 3 (first century A.D.) were found here: both are now in the Museum at Parma. One would imagine that this is the site of the discovery by Ficker of ib. 1202. 2, 1265 b. 2 (both first century), 1292. 3 (123 A.D.), recorded as in parietinis rnagnis meridiem versus. To the N.W. of the main substructions (at a in our plan) is a great mass of concrete, about eight metres square, with niches, which Grossi- Gondi (tav. ix.), I think rightly, considers to be a tomb, while Canina brings it well within the area of the building. I think this should rather begin at a', where there is a low mass of concrete. Rossini shows a circular tomb a little to the N.W. Opposite to it on the S.W. and N.W. (at 3, 3' on our plan) are remains of buildings, well figured in Rossini's plan and also shown by Canina—tombs (some columbaria), no longer clearly visible.3 In 1807 Lucien Bonaparte is said to have excavated a group of tombs on the ridge above the Rufinella, where Tusculum faces the Alban Mount, on the right (S.) of the ancient road, i.e. not far from the amphitheatre either on the E. or W. The building is described as though it had been a house, but was in reality of a sepulchral character. It was approached by some peperino steps: one of the rooms with a mosaic pavement (with a bust of a youth in the centre holding a lituus) contained four arched niches, under which bodies were buried, enclosed in slabs of marble, while at each angle was a wine jar full of burnt bones. The unimportant sepulchral inscription C.I.L. xiv. 2760 was found there, and a brickstamp (ib. xv. 1039, 2) of 123 A.D. was also discovered in one of the rooms. The whole is described at great length, but not very clearly, by Guattini in Memorie Encicl. iii. 130 ; but from a comparison with Rossini's plan it seems clear that the room he is describing is that nearest to the 1 Ib. 809, a. 5 is recorded as having been copied by Henzen ' in the ruins called the Villa of Cicero.' I am inclined to suspect that the site is the same ; and agree with Grossi-Gondi in referring to it also ib. 2226 (1st cent. A.JJ.). 2 One of these is now in the Villa Aldobrandini : it is just possible that ib. 1029, a. 12 (now in the same place) was also found in these excavations. 3 The sepulchral inscription found in 1861 in the remains of the great building of which we have spoken {C.I.L. xiv. 2679) must belong to the tombs of this part of the road. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 347 amphitheatre in the northern group 3'. Guattani is in this case placing the tomb on the wrong side of the road. To the N.W. of the building numbered 2 on the left edge of the road, ascending to the N.E. (the pavement of which is well preserved) I have noticed three fragmentary inscriptions: (1) [C TOAE, on a fragment of a white marble epistyle, O"26 metre high, the letters being crio5 metre high (which I saw some years back, and have not been able to find again). (2) C.I.L. xiv. 2645, on the narrow edge (C23 metre wide, the letters being 010 high and well cut) of a block of marble 078 by C54 metre, with bolt holes on the upper surface. The second full point should, I think, be omitted. (3) COCTVS OL C in rough letters croo, metre high, on the narrow edge of a block 1*27 by 06 metre. They come very possibly from these tombs, and (1) and (3) do not seem to be published elsewhere. At the N. angle of the large building the main road turns S.E., but a branch goes off due E., which Canina calls the 'Via Particolare delle Ville.' Following this we reach, at the point marked 4 in our plan, the site of the house, with a small atrium and a reservoir (of which he gives a plan on tav. xxiv.: cf. p. 132) which was excavated in 1827 {Edifizi, VI. tav. xciv.). Of the house no remains are now visible, but the reservoir is still to be seen. The plan is also given by Angelini and Fea (tav. xi.), and their rendering is probably more correct, inasmuch as there seems to be some amount of arbitrary restoration in Canina's work. Below it are other ruins, 4', now no longer visible, which Canina thought to be the main building of the villa; in them were found in 1825 a putto with a dog or hare, a double Bacchic herm, a head of Jupiten a marble medallion with a representation of Apollo, a bust of an Indian Bacchus, a bust of a female member of the gens Rutilia (this statement is not consistent with that of p. 143, infra, 352), a head and a bust of Diana, and a fragment of a finely carved group of Bacchus with two Fauns (Canina, pp. 133, 149, and tav. xxxvii., xxxviii.). In regard to this last we have a curious application from Biondi in the Atti del Camerlengato, Tit. iv. fasc. 1291. In a letter of March 6th, 1830, he suggested that, as in 18261 two torsi and fragments of a third statue, of Greek marble and well carved, forming a group of a Bacchus dancing between two fauns, had been found at Tusculum, while in the previous excavations of Lucien Bonaparte two moderately good statues

1 The exact date was April 25th. (Canina p. 144 : on p. 8 he wrongly gives the date as 1830.) 348 THE BRITISH SCHOOL AT ROME. representing ladies of the gens Mamilia had been found (which were included in the purchase of the Pallas Giustiniani and other statues in the Museo Chiaramonti, and were then in the Vatican storerooms), and, as these last would complete the royal collection at Agli^ near Turin, he would wish an exchange. The Commission, however, reported against it, adding that the Mamiliae were really Rutiliae, and the exchange was refused. The group was then sold to the King of Prussia, and is now at Berlin {Beschreibung, no. 96). Canina gives two plates of it (Tuscolo, xxxii., xxxiv.), one representing it before its restoration, the other an attempt at a restoration by Bisetti, whereas this work was actually done by Wolff. All the rest of the sculptures were removed to Aglie. Some ornaments in terracotta are given on tav. xlix.1 On the S. side of this road are two other buildings, 5, 6, at divergent orientations, called by Canina ' Casa del Console C. Prastina Pacato.' (See also Rossini, op. cit., for the plan.) Here, then, was found in 1825, the base bearing the inscription C.I.L. xiv. 2588, probably a dedication to Fortune made by the consul ordinarius of 147 A.D. Biondi (MS. f. 4) adds to the discoveries in this house (or in the ruins below it on the N. of the road, for his account is vague) the fragment of a calendar C.I.L. xiv. 2575, and the lead pipe with the inscription Reipublicae Tusculanoruni {C.I.L. xv. 7813 : cf. infra, 357). Specimens of the paintings from this house are given by Canina (tav. xliii. = Edifizi, VI. 96). They were transported to the castle of Aglie. The remains now visible are scanty. At 6 there is a long wall going N.E. with substructions in opus reticulatum of tufa. Beyond these ruins the ancient road bifurcated: the right hand branch led to a small open space, paved with slabs of sperone, with seats of the same material, which stood in front of another villa (7 on our plan) of which scanty traces are visible, where it ended. To this villa, excavated in October, 1838 (Rossini indicates only a few walls on the site), was given the name of the house of the Caecilii, from the discovery in it of a painting, which was supposed to represent a scene in the life of this family, Biondi notes that, according to tradition, Tanaquil, the wife of Tarquinius Priscus, when she came to Rome, acquired the name of Caia Caecilia, and was celebrated both as a spinner of wool and as an interpreter of prodigies (cf. the passages cited by him in Diss. Pont. Ace.

1 I may notice in passing that AM cit. fasc. 1098 deals entirely with the exportation of modern copies in plaster and marble from the Runnella to Palazzo Chablais and thence abroad. CLASSICAL TOPOGRAPHY OB- THE ROMAN CAMPAGNA.—III. 349

Arch. X. 386 sgg.) A priestess or diviner is, according to him, represented by the old woman seated in a pensive attitude on the right of the picture : on the left sits a younger woman crowned with laurel, and in the centre is a girl spinning, about whose destiny in marriage an oracular response is being demanded. He refers the painting to the incident related by Cicero (De Div. 104) and Valerius Maximus (iv. 4). Caecilia, wife of a Caecilius Metellus, whose identity cannot be precisely fixed (not necessarily Balearicus, as Biondi thinks), though he was a contemporary of L. Valerius Flaccus, consul in 100 B.C., went with her sister's daughter to consult an oracle in regard to the marriage of the latter. The former sat, but the latter stood, and becoming tired, asked her to let her rest on her seat: whereupon the aunt said ' I willingly give up my place to you'; and so it fell out in reality: for the aunt soon died, and the niece married the aunt's husband. Biondi gives a plate only of this incident, but describes other paintings also. Canina gives the plan (tav. xxiv. fig. 1 = Edifizi, VI. tav. xciv) and representations in colours (tav. xli., xlii.—not repeated in Edifizi) of this painting and another, and (tav. xlvi^-xlviii.) of one of the marble pavements and of some of the decorations in stucco. In one room a number of small pieces of marble for use in a pavement (which Biondi thinks had never actually been used) were found ; while in others there were marble and mosaic pavements. The excavations were continued at the end of 1839 after the death of Biondi, under the direction of Canina, as he tells us in his text (pp. 81, 133) and a marble vase adorned with groups of griffins and stags was found (p. 151, tav. xxxix., xl.), and also a head of a Roman lady very skilfully painted on plaster (p. 149, tav. xxxviii., fig. 4). Two of the brickstamps give the dates 123 and 124 A.JD. (C.I.L. xv. 270 a. 7, 2025. 1) while 2280 seems to belong to the middle of the first century A.D. A wall of opus reticulatum with brick bands is still standing at the N.E. angle (7'). The left hand branch of the path retains its pavement for a short way only, and after a time turns N. passing a little E. (but not so near as Canina represents it) of a circular tomb (8 on our plan) called La Torrex (Tusculo, p. 136 and tav. xxvi.: cf. Edifisii VI. tav. lxxxiv.) constructed in opus

1 In this work a restored sectional elevation is given, and a plan, which is not really that of La Torre, but that of the so-called tomb of Lucullus, repeated from the preceding plate. 2 xlvii, xlviii are repeated in Edifizi vi, tav. xcix. 350 THE BRITISH SCHOOL AT ROME. reticidatum: it is circular, with a chamber in the form of a Greek cross. With the further course of the road to the N. we shall deal later: it has a branch going E. to join the path up from Camaldoli {infra, 371) passing the remains of a building (9 on our plan). We now return to the temple (?) no. 2 on our plan, and follow the paved road leading due E. from it. The modern path does not coincide with this exactly, but runs a trifle further S., following the boundary wall between the two properties already mentioned. We very soon reach the point at which Canina marks the site of one of the gates of the ancient city (10). What remains here is simply a mass of concrete vaulting which has fallen, and it can of course have nothing to do with the original fortifications of the city, of which as a matter of fact no traces remain, nor does Canina maintain that they do. The scanty traces of concrete along its supposed course, e.g. at 11, naturally belong to the imperial period; though it is probable that Canina is fairly right (p. TJ and tav. \\. — Edifizi, VI. tav. lxxxv.) in determining the line of the confines of the ancient city by the fall of the ground. As a fact everywhere except on the W. their line can hardly be conceived to have differed from that given by him, as the ground falls away steeply on all sides. The western termination, on the other hand, might more appropriately be placed at the amphitheatre, where the neck of land is a narrow one (cf. also supra, 341). Whether the piece of walling near the fountain (no. 20) belongs to the early city wall we shall inquire later [infra, 360). After a little over 200 metres, the road divides, turning to the N.E. and S.E., to avoid the Forum (13), an open space in front of the theatre sur- rounded by colonnades ; just at the turn are the foundations which Canina considers to be those of the Curia (12) and of which nothing is now visible ;1 then the road runs E. along its N. and S. sides again, though the pavement is preserved only along the N. side. I cannot offer an opinion as to the correctness of the lines of the cross streets which Canina indicates in his plan, nor of the details of the Forum which he gives (Tuscolo, p. 116 sqq. and tav. x^Edifizi. VI. tav. 86); the cornices, etc., which belonged to a portico with Corinthian columns, were conveyed to the casino of the Villa Rufinella, having been found in the excavations of Lucien Bonaparte

1 This must be the ' temple facing S. with fluted columns of Luna marble with Corinthian capitals ' of which Biondi speaks (MS. f. 9). CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 351 early in the nineteenth century (cf. tav. xvii, figs. 7, 8), but Canina claims to have been the first to recover the plan. A capital in sperone in imitation of the Egyptian style is also given ibid. fig. 9 ; but the exact site of its dis- covery is unknown. Biondi describes (MS. f. 9/ sqg.) at the precise place where Canina places the Forum an entrance door on the S. of the road, leading to ' a public or at least a noble building': on each side of it was a base with the impression of a column upon it, one of the drums of which was found and re-erected. Beyond this was the threshold of a shop in sperone. Further on the left was the approach to an almost circular temple, which one descended by steps (not shown by Canina), which had been converted into a Christian church and roughly restored in the Middle Ages. Indeed along the road there were evident traces of a cemetery with many skeletons, some lying on the pavement of the road, others in rough coffins of walling or of stone. When these were removed, the road was found to continue, and soon had a branch to the right. This was not followed, and, continu- ing along the straight road, there was found to be a wall in opus quadratum of sperone, restored in opus reticulatum, on each side of it: there were traces of the impost of an arch which passed, not straight, but diagonally across the road (this is the lateral passage on the N. side of the theatre under part of the cavea). The road now began to ascend, and at the beginning were some houses with walls painted red. Ascending more rapidly the road then left on the right the cistern no. 16 which had been found by Lucien Bonaparte and cleared again by Biondi, who noted that its sides were not perfectly parallel. The first part of the account hardly seems consistent in detail with Canina's description : it would be very desirable that the remains should be again uncovered, so that we might learn the exact truth. But from the position of this area, which is precisely identical with that of the Forum at Ostia, and from the discoveries made in or near it by Lucien Bonaparte, it seems almost certain that this was the Forum, and Dessau's scepticism {C.I.L. xiv. p. 258) is a little excessive. According to Canina's statements (though unfortunately accurate notes of the exact site of the discovery of the various objects were not taken) fragments of a statue of Augustus, which fitted on to a torso already discovered, were found near the Forum on Oct. 16th, 1825 (p. 140, tav. xxx : cf. Biondi MS. 352 THE BRITISH SCHOOL AT ROME.

f. 4) and removed to Aglie. Two toga statues were found in the same locality by Lucien Bonaparte : one bears on its base the name of Cn. Velineius (C./.L. xiv. 2754), while Biondi (MS. 77 sqq.) reports on Oct. 9th, 1825 the discovery of some fragments which completed the other. Canina figures both in Tusculo tav. xxxix; they were conveyed to Aglie (though there seems to be some doubt as to whether they are still there—Dessau in C.I.L. xiv. p. 255, n. 4). Among the objects found hereabouts by Lucien Bonaparte were a statue of Antonia, now in the Braccio Nuovo of the Vatican (no. 77), two female statues now in the Museo Chiaramonti (35Si 357) with inscriptions on their bases, showing that they represent two members of the gens Rutilia {C.I.L. 2741, 2742 : cf. Canina, tav. xxxii: the former inscription is probably best restored, as Hiilsen thinks, Rutilia L.f., mater Ter{entii) Reginii)—cf. Amelung, Sculpt, des Vatik. Museums, i. p. 545), and a headless bronze statue (the whereabouts of which is not known).2 A bust supposed to represent another female member of the gens Rutilia was found in 1825, where the statues had been found, and conveyed to Aglie (Canina, p. 143, tav. xxxvii), but cf. supra, 347. Several other sculptures found by Lucien Bonaparte are figured on tav. xxxviii and described on pp. 149, 150. They include a head of the youthful Augustus, another head of Corbulo, three unidentified heads, a head of the youthful Hercules or Theseus, and two marble disks with Bacchic reliefs. They were to be seen in the palace in Via Condotti which he occupied, but I do not know where they are now. The objects found to the S. of the Forum, in the tenuta della Molara in 1829, 1830, and 1834 are mentioned supra, 344. Bases with inscriptions indicating the existence of other statues were found by Lucien Bonaparte, some of historical personages {C.I.L. xiv. 2600, Q. Caecilius Metellus cos., 2601, M. Fulvius M.f. Ser. n. cos. Aetolia{in) cepit—the latter was consul in B.C. 188, but the former cannot be certainly identified), others of characters in Greek mythology (Orestes, Pylades, Telegonus,3 Telemachus, 1 He mentions that it had been placed (at the Villa Rufinella) on the base bearing the inscription C.I.L. xiv. 2753, which did not belong to it. 2 I suppose this is the fine Apollo in bronze slightly less than natural size mentioned by Nibby (Analisi, iii. 328) as still in his possession. The 1818 edition of Vasi's guide (Itinerario di Roma, p. 221) enumerates among the objects found at Tusculum and existing in the Palazzo Bonaparte in the Via Condotti a marble vase, the Apollo, four busts (one supposed to represent Germanicus, the other Perseus, which were much praised), and the two Rutiliae. From the construction of the sentence it is not clear whether the statue of Tiberius mentioned just before them should be included, but probably not. 3 Above this name we see Iaso (n); but Mommsen is undoubtedly right in supposing that the CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 353 and the poet Diphilus, ib. 2647-2651). Canina states that the former group stood in the Forum, the latter in the theatre ; but Dessau thinks that as the bases are all of sperone, and date from about the time of Augustus, they must all have been found together. Certainly the last group is more appropriate to the theatre, but I think Dessau goes too far in regarding Canina's Forum as entirely fictitious. One may well suppose that the inscriptions built into the small modern house on the supposed site of the Forum were found there or thereabouts. They are as follows: C.I.L. xiv. 2611 (the meaning of which is very obscure) ; 2623 (fragments of a large epistyle about C65 metre high, with an identical or similar inscription on each side of it, relating to the construction by [C. Caejlius C. f. Rufus of steps round an area and an out- let channel (emissariuni) of travertine for games1) ; 2632 (fragments of a slightly smaller epistyle). Parts of a similar epistyle (ibid. 2622) with the names of C. Caelius Rufus and C. Caninius C. f. Rebilus (the latter perhaps the consul—for one day only—of B.C. 45, or his son or grandson) is at the Villa Rufinella. A sepulchral inscription (ib. 2746) is also built in here. A view of the area probably occupied by the Forum, looking west- wards, is given in PI. XXIX. Fig. 2. On the E. side of this is the theatre (14) (PI. XXX. Fig. 1), the stage of which was entirely brought to light under the direction of Canina in 1839-40, the excavations of Lucien Bonaparte having led only to the uncovering of some of the seats in the lower part of the cavea, as may be seen from the plans of Angelini and Fea (tav. xii.), and Uggeri (tav. x.). The cavea was built against the hillside, and was not supported by substructions :2 the stage still has the chamber beneath it, reached by rectangular shafts, from which the curtain was drawn up, and the whole is well preserved as far as the plan goes, though not above ground level. Details will be found in Canina, Tuscolo, pp. 118 sqq. and tav. xi.-xiii. ( = Edifizi, VI. tav. Ixxxvi., lxxxvii.) : cf. p. 127 and tav. xvii. ( — Edifizi, tav. lxxxix.), where he figures Ionic and Roman Doric capitals and bases of sperone as having been found in the theatre, whereas on p. 118 he says upper name was cancelled, and that of Telegonus, the founder of Tusculum, substituted ; and two statues could not have occupied the same base. 1 This lettering belongs to the time of Augustus in Henzen's opinion, and so the inscription cannot be associated with the amphitheatre—nor is it at all likely that it would have been transported thence to this place. 2 Nibby speaks of a drain running round under the lowest seat existing {Schede, i. 64), which is not now visible. A A 354 THE BRITISH SCHOOL AT ROME. that the back wall of the stage was decorated with Ionic and Corinthian columns. In any case these were of course coated with stucco. No inscriptions relating to its structure were found in it, except a fragment with the word Magister (so Canina ; but the inscription has been read differently—C.I.L. xiv. 2641), but a cuirassed statue of Tiberius was discovered {Tusculo, pp. 8, 138, tav. xxix.), the head of which had already been found by Lucien Bonaparte, and the whole was carefully restored by Cauda. I do not find it mentioned by Bernoulli in his Rom. Ikonographie, and where it is now I do not know. Canina does not note it as having been conveyed to Aglie. To the S. of it are three or four steps on a curve (15), which have by some been believed to be an Odeon or smaller theatre ; for this, however, the space is far too small, and Canina (pp. 114, 117) rightly pronounced them to be simply a fountain fed by the large square reservoir (16), the roof of which was supported by three rows of five pillars, which is to be seen behind the theatre. It was excavated by Lucien Bonaparte and its plan will be found in both the works cited, Canina repeating it on tav. ix. (cf. Edifizi,N\. tav. lxxxi.,lxxxvi.) for comparison with other cisterns— those of the so-called Villa of Galba {supra, 312) and that of the Villa Taverna {infra, 371). It must have been supplied mainly by rain water, but also perhaps from the water collected in passages under the arx (Canina, p. 79). To the N. of this large reservoir is another, triangular in shape (17), and to the N. of that again the remains of a small house with an atrium with four columns are shown by Canina (18), but not by Rossini.1 There is no pavement visible in the ascent to the arx where Canina marks it; but on the way up, at the point marked in the map, there is a cave, in one corner of which is a small cistern cut in the rock, which I think is not ancient: above it are some small niches cut in the rock and above that a chamber. One wonders whether it can possibly have been a hermitage of the Middle Ages. Upon the summit of the hill, where Canina is undoubtedly right in placing the arx,2 there are no remains of even the classical period : every- 1 I imagine that this is the house referred to in Bull. Inst. 1829, 213, 'On the road already excavated, with a gate of the ancient city, the ruins of a private house were found, containing many votive terracottas and a statuette (idolo) of bronze. These ruins were considered to be those of a house, because of the atrium, portico, and impluvium found there, but what is really remarkable in this discovery is, that close to this house another paved road is said to have been found, about 12 palms (2"67 metres) below that already discovered, which is also ancient . . . we await from Marchese Biondi the important results of his work.' 2 The arx is mentioned in the various accounts of the attacks on the town, e.g. when the arx CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 355 thing that I have seen there seems to belong to the mediaeval buildings and fortifications destroyed in 1191, and I could distinguish no traces either of temples or of ancient city walls. Canina himself {Tuscolo, p. 75) states that in the excavations of 1835 and 1836, which he carried on by order of Prince Borghese (within whose property the arx fell), he found only buildings in which older material had been used, e.g. sarcophagi which served for later burials ; and the squared blocks of stone noticed by Nibby {Analisi, iii. 320) are not necessarily ancient. Nibby and Canina's south gate of the citadel {Tuscolo, p. 112 and tav. viii.) is merely a gap in the rock, where the foundations of a house or a tower had been formed, by cutting the rock away so thin that it has at last given way (d). There are considerable traces of similar cuttings in the rock on the S. and E. sides. The only piece of Roman construction I have been able to detect is a substruction wall of Roman concrete at the N.E. angle below the top (19) forming a right angle. For a view of the arx see PI. XXXI. Fig. 1. Among the blocks of stone which support the cross (which were placed there in October, 1864) an important inscription (now in the Villa Aldobrandini) was recognized in 1901, and published by Grossi-Gondi (// tempio di Castore e Pollnce nell' Acropoli di Tuscolo, Rome, 1901) and by Tomassetti {Not. Scavi, 1901, 280). It is cut upon a cippus of sperone O'6g metre high and runs thus : ae \a\edis Ca[s(toris) ei\ Pol{hicis) ex d{ecurionurn) d{ecreto) M.' Avillius D. I. Stabilio . . . Anicius P. /..... Furius P. I Plaetorius D. [/.] . . . Volcacius C. I. mag{istri) fac{iundnni) coer{avemnt). The inscription seems to belong to 60 or 70 B.C. The restoration of the first word is not certain : Grossi-Gondi proposes aedi{culum), aetoma, or derar{iuni), though he thinks that the second is less probable because there is a vertical stroke after e, which is too close to it to belong to a t. Tomassetti proposes aeditui, which seems to me impossible. Grossi-Gondi maintains that it is highly improbable that the cippus was brought up here for the erection of the cross : there I agree with him (and the contemporary testimony which he cites seems to show that the stones were collected on the spot), though I do not think we can be certain that it was not moved in the Middle Ages. Still, there is no doubt that this is a very likely site for the Temple of Castor and Pollux, whom we know to have been {infra, 362) itself was for some time occupied by the Aequi in 459 B.C. (Liv. iii. 23) and when the whole city except the arx was surprised by the Latins in 377 B.C. {it. vi. 33). A A 2 356 THE BRITISH SCHOOL AT ROME. the tutelary deities of Tusculum, cf. Albert, Le Culte de Castor et Pollux en Italie {Bibliotkeque des Ecoles francaises, fasc. 31), p. 13: that its sub- structions are not recognizable may surprise us (for I cannot agree with Canina in seeing traces either of this temple or of that of Jupiter, both of which he wishes to place on the arx), but we must put that down to the transformations of the Middle Ages and the total destruction of 1191, and remember the fate of the Temple of Jupiter on Monte Cavo {infra, 398). For a view of Monte Cavo from Tusculum see PL XXXI. Fig. 2. The allusion in Eschinardo's Esposizione, as revised by Venuti (p. 268), to a ' muraglia altissima, parte fatta con scalpelli in quel sasso, e parte aiutata di sostruzione' between the ridge and the highest summit, is merely a confused description of the wall by the fountain and the rocks under the highest summit with the mediaeval defences, and should not be taken to allude to the existence of any considerable stretch of wall not now visible. The passage does not occur in the original edition. At the N.W. angle of the Forum the descent towards Camaldoli commences and here we reach at once an interesting group of monuments (plan 20), the discovery of which Canina (p. 123 : cf. his plates xiv.-xvi. ( = Edifizi, VI. tav. Ixxxviii., lxxxix.) and the far better representations of the state of the work, as far as it had gone in 1826, in Rossini, op. cit. tav. 72, 73) is wrong in attributing entirely to Biondi's excavations in 1825, whereas they were in reality partly laid bare by Lucien Bonaparte and are described by Nibby, Viaggio Antiquario (1819), ii. 47, sqq. Their respective shares in the work are clear from Biondi's report, which is to be found in the Atti del Camerlengato, Tit. iv. fasc. 2461. The site of an entrance to the city may reasonably be placed here ; but Canina (p. 126) expressly warns us that only the foundations of what may have been its piers were found, and that the drums of half-columns shown in his tav. xvi. were only placed there to indicate its position. We first see, on the S. (right hand) side of the road a piece of embankment wall of opus quadratum of porous tufa : the lateral joints are

1 A passage in Nibby's Schede (i. 103) shows that Biondi's work began as early as 1817 : 'in the excavations made by order of the Court of Turin, to which it belongs, under the direction of Marchese Biondi, there have been so far discovered (September 1st, 1817) another piece of the road which runs along the walls; and after leaving on the right a small hemicycle in reticulatum for the repose of travellers, it divides, and one part runs to the left and joins that which ascends from Camaldoli, the other to the right joins that which comes from the Villa Rufinella : here there seems to have been one of the gates.' CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 357 not vertical in all cases, and there is some anathyrosis ; but the horizontal joints are regular. The height of the courses averages 035 metre : there is no alternation of headers and stretchers. Some of the latter are very long : one block I measured runs to 1'94 metre. In the lower part of this wall is a rectangular opening rabbeted for a door (at b : cf. Plate XXX. Fig. 2) which leads into a chamber 2-85 metres long and 3'10 wide at the ground level : the latter measurement decreases owing to the gradual inclination of the sides ; when these have approached sufficiently the ceiling is formed by two slabs meeting to form a pointed roof.1 The courses in the wall opposite the door measure, on an average of ten blocks, C275 metre high. Some of the joints are not vertical, and in one case one block is cut out to allow of the angle of another entering it, the level of the course changing slightly. From this chamber a narrow passage, at first built of opus quadratum but afterwards cut in the rock, runs into the hill, passing, as indicated in Canina's plan (tav. vi.) under the theatre and the large reservoir to the E. of it, and dividing into various branches when it arrives under the cliffs on the W. side of the arx. These passages still serve for the collection of water, and, being of about a man's height, can be followed without much difficulty. Within the chamber there were traces of arrangements for filtering, or more probably for dividing the water into three portions, one of which fed a small fountain some 4 metres further down the slope : the basin, a rect- angular trough of red tufa, no by 1 metre, still bears on the front, in letters C05 metre high, the inscription [Q. Coelius] Q. F. Latin{iensis) M. Decumiu{s) aed[iles) de S(enatus) siententia) {C.I.L. xiv. 2626). The words Q. Coelius have been obliterated, since the inscription was found, by a fracture of the stone. Dessau thinks that the first of the two aediles may be identical with the man of the same name mentioned by Cicero in De imp. On. Pompei 19 § 58 (i.e. a little before 66 B.C.) as tri'bune of the plebs and legatus of a proconsul in the year following ; and the lettering of the inscrip- tion (figured by Ritschl (P.L.M.E. pi. L.c) from Henzen's copy) would warrant this supposition : within it is a smaller square block of stone with a round basin cut in it. According to Canina, the lead pipe with the inscription C.I.L xv. 7813, Reipubilicae) Tusculanorum was found here, and served for the supply 1 Two of the slabs near the door have fallen, and the roof has been mended (in modern times ?) with concrete. 35^ THE BRITISH SCHOOL AT ROME. to this fountain.1 That it was discovered in 1825 there is no doubt, but in the list of objects sent to the Camerlengato by Biondi, it figures among those found outside the excavation,2 which is defined to have taken place

1 Ibid. 7833 Cabarasiae P. f. Pacatillae, xii was found in the excavations of Lucien Bonaparte. 2 It may be well to give here a summary of the documents in Atti del Camerlengalo, Tit. iv. fasc. 246. On May 21st, 1825, Luigi Biondi, on behalf of the King of Sardinia, applied for leave to export the objects which might be discovered on his property, as it had been given to other sovereigns. The Cardinal Chamberlain opposed the demand, noting that the other precedents did not hold, as other sovereigns had transported objects from their own palaces, and that the Barberini Faun had only been allowed to go out after two years of discussion by express order of Pius VIII. On the 13th of June the Cardinal Chamberlain wrote to Biondi to come and interview him, and apparently permission was given as a result. On July 4th Biondi wrote that the King's intention was to excavate in the Villa Rufinella towards Camaldoli. Rossignani, who was sent to examine the site, reported on the 12th that it was just where Lucien Bonaparte had brought to light a portion of the Via Latina, the so-called Arco Acuto (the chamber described above), the aqueduct, and part of the city walls, and permission was granted on the 16th. Work began in August, and on October 24th it was found necessary to remind Biondi that he must periodically send reports of his discoveries. Biondi replied that at first nothing had been found, and that he had believed it his duty to inform the King first. (In his MS. indeed, we have a series of eleven weekly reports from September nth (?) to November 2cth, which were, I presume, those sent to the King; but they do not give any precise local indications.) There follows, as a result no doubt of the remonstrances made, a sheet, in the hand of Biondi's secretary, with a list of the objects found, without date. A duplicate of it will be found in Biondi's MS. f. 101. It is practically a summary of the monthly reports. The list is as follows :— A. Outside the excavation (C.I.L. xiv. 2575, 2588, 2636). Fragment of a sepulchral inscription of several freedmen (ib. 2693 ?). Lead pipes with the stamp Reip. Tusculanorum (C.I./,., xv. 7813 ; ibid. 7860 and 7876 seem to have been found at the same time). Fragments of paintings found in the ruins (these are described in the weekly reports and may be recognized on tav. xliii of Canina's Tuscolo—supra, 348). Fragments of terracottas (also described in detail in the reports). Three small columns of gray marble, only one entire (described in the weekly reports as of cipollino, \\ palm (C33 metre) in diameter and 10J (2'34 metres) long). Four weights (circular). A rabbit in marble of inferior style (eating grapes). Fragments of a candelabrum.. A small capital of a column. A tazza of pavonazzetto in several pieces, with almost the half wanting (with a flower in the centre). Head of Jupiter, and two other weathered heads (including (?) a female bead of natural size mentioned in the report for October 22nd with the Jupiter). A headless bust. Small fragments of statues, of which some belong to statues found by the Prince of Canino. B. In the excavations (quoted in full in the text). From the weekly reports we may add (besides the objects mentioned by Canina—cf. supra, 351) a cameo from a ring, representing a head of Bacchus crowned with white and black grapes, fragments of stucco, a head of Antoninus Pius and another unknown male head, an ex voto in marble (a disk with on one side a sea-scene with dolphins and on the other two heads) and a small bas relief of Jupiter Ammon. The additional details which I have gathered from the weekly reports about items already in the list have been added in brackets. On December 12th, 1825, Biondi announced his intention of closing the excavations, at least for the present, because the season was unfavourable, and the results less good than was expected : CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 359

'just where Lucien Bonaparte had brought to light a portion of the Via Latina (sic), the so-called Arco Acuto, the aqueduct, and part of the city walls' exactly here ; whereas within its area were discovered, according to the list, ' the continuation of the external road found by him, a milestone he said that lately a small sepulchral inscription of a freedman (C.I.L. xiv. 2671—cf. infra, 360n.) had been found, and a fragment of a bas-relief, and also a fragment of a ' putto.' The road which had been found, he added, ran past the theatre and the large reservoir towards the citadel. Rossini, op. cit. tav. 70-72, gives an excellent idea of what had been found up to this time, both as to the state of the excavations along this piece of the road and as to the fragments discovered, though he unluckily mixes up what was already at the Villa Rufinella and what had been more recently found. On March 31st, 1826, Biondi gave notice of the resumption of the excavations ; and the Cardinal Chamberlain advised Carlo Fea, as Commissario delle Antichita, and the Governor of Frascati in the usual way. On April 30th, Biondi reported that on the previous Tuesday two torsi of statues belonging to the same group had been found, they were miserably ruined and defective, but of good style : the governor of Frascati wrote on May 3rd, that one seemed to be a Satyr, the other a Bacchus (the Berlin group). On August 21st Biondi reported the discovery of a seated statue larger than life-size, lacking the head, the left arm, the right hand, and one leg (the so-called Tiberius, supra, 344). After this we have nothing, until Biondi wrote on April 12th, 1828 of his intention to resume work : in reply to which he was requested to send the permit of excavation for extension, as it had expired over a year before ; and a note states that on May 9th it was renewed till the end of the year. On July 22nd, 1829 Biondi applied for a further extension, and on September 27th, 1829 we get a report that during the previous week the excavations had been resumed, and the following objects found : two fragments of a marble candelabrum with leaves and fruits ; a small statuette in bronze ; many votive terracottas, representing heads, arms, legs, feet, small oxen, small pigs, etc. Nothing more was heard of Biondi (who did not personally send this report) and on April 21st, 1830, it was decided to write again, as a note on the back of the document says, but apparently without result. The excavations were not resumed till the autumn of 1839, after his death. The next document belongs to October 21st, 1839 and is a letter from the Governor of Frascati mention- ing the discovery of a finely carved torso, believed to be of .Greek workmanship. On November 2nd, 1839, came a report by Canina, who was now in charge, on the excavations of the last week in October. A torso of an imperial statue was found, of fine sculpture, but lacking the head, the arms, and one leg (the Tiberius, supra, 354) ; and the stage of the theatre, with the stairs up from the orchestra, and the slit for the curtain, was uncovered. On December 1st Canina reported that in the last half of November a white marble vase decorated with hippogriffs and stags {supra, 349) had been found, and that the work in the theatre continued. In November, 1840, Canina applied on behalf of the Queen fry leave to resume operations at Tusculum and at Isola Farnese, which was granted. Of the results of the work of 1840 and 1841 there are no accounts ; and the next paper is Canina's application of August 20th, 1842 for a renewal of permission, which was granted. On November 6th Canina reports that in the remains of an ancient villa various fragments of paintings with ornaments and figures had been found, which belonged to vaulting now entirely ruined, also various fragments of terracotta. A statue of a Bacchic figure was also discovered, in small fragments and incomplete. In the same place two small round vases were found with inscriptions relating to a military tribune of the Gens Furia, whose tomb was not far off {C.I.L. xiv. 2577, 2578—infra, 373). Lately, too, there had been found a well preserved chimaeia in marble with a good statue, rather larger than life-s:ze, representing Minerva, but wanting the head (which had been in a separate piece) and one arm. The objects found were deposited at the Villa Rufinella. This is the last document we have on the subject. 360 THE BRITISH SCHOOL AT ROME. with the number xv, the entrance to a vault, a part of a wall, and the place where the city gate was situated ; and the internal road, dividing into two branches, one going towards the theatre, the other towards Frascati.' On the other hand in Biondi's first three weekly reports it is described as end- ing at a small fountain : it is added that it was 70 palms (15*56 metres long) and about 2^ oncie (o'OS metre) in diameter, so that Canina is probably right; and this furnishes another ground for refusing to attribute the whole group of constructions to an early date. The ' milestone with the number xv,' a small column of sperone, was found quite close to the fountain (cf. Tuscolo, tav. vi.) and has been seen by no one since Canina's day ; but I do not share Dessau's doubts {C.I.L. xiv. 4088) as to its having been a milestone at all. For the distance measured from the 9th mile of the Via Labicana is almost exactly 15 miles (even from the 10th mile of the Via Latina it is over 14) and Dessau in taking the rough estimates of Dionysius and Josephus (about 100 stadia) au pied de la lettre, is falling into the same error as Grossi-Gondi {supra, 317). There seems to me to be considerable doubt whether the embankment wall and the chamber with the pointed roof are in reality of very remote anti- quity. The construction in other respects does not point to it, and the wall seems much more like an embankment wall than a massive fortifica- tion wall. The pointed architecture is of course curious, but I am inclined very much to doubt whether it points to an actual ignorance of the principle of the arch. Canina, following Biondi, conjectures (p. 126) that an arched aperture close to the milestone may have led to the baths (though there is little reason for his supposition) ; but this, according to Biondi's notes, was of opus quadratum, and the fallen pillars of the arch were largely restored by him (f. 6V). It is now buried some 50 yards further down ; on the same side is a rough concrete vault. On the opposite side of the road Nibby notes the existence of tombs1 and of substructions to the road (for the latter cf. Analisi, iii. 329) in opus reticulatum ; and further to the N. are massive substructions (plan 21) in concrete. Canina figures (tav. xvii. figs. 1, 2, cf. p. 127) a Doric cornice and base in sperone (' pietra tuscolana,' as he calls it), fragments of which were found along the road, and which, he conjectures, belonged to some building above the substruction wall. The pavement is well preserved, and

1 To these tombs one may attribute C.I.L. xiv. 2671, found in 1825 'outside the city' by Biondi. CLASSICAL TOPOGRAP'IIV OK THE ROMAN CAMPAGNA.—III. 361 the road descends to Camaldoli, but we will not at present follow its course further (see infra, 371). To the S. of the modern boundary, within the Tenuta della Molara, on the S. slope, there are considerable remains (22, 22', 23) of substructions. At 22 and 22' only foundations in concrete are traceable, but at 23 there are some large vaults, partly in opus reticulatum, running S.W. and N.E. belonging no doubt to a building facing S.W. Below there is a small rectangular reservoir, and a substruction some 20 metres in diameter, in which is a curved corridor about 5 metres wide, which once perhaps formed a full circle : it is faced with rough pieces of selce, and is mentioned by Nibby, Schede, i. 103. It is strange that these remains are not indicated at all by Canina in tav. vi. and only inadequately in tav. vii. On a projecting and somewhat lower spur (24) are traces of buildings consisting of small rectangular chambers close together and marked by Canina as tombs (no doubt from the evidence of actual discoveries, for otherwise they might be dwellings) ; they, too, belong to the same period ; while at £ is a well or quarry shaft. As Dessau remarks in his preface {C.I.L. xiv. p. 255), it is very often extremely difficult to find out the exact spot at which were discovered even those inscriptions which we know to have been found on the hill of Tusculum in the first half of the nineteenth century, for many of them were at once transferred to the Villa Rufinella and copied there. Among the more important I may mention C.I.L. xiv. 2579, a dedication to Jupiter and Libertas, erected during the aedileship of P. Valerius Bassus praefectus fabrumv) 2581 (a dedication to Mars Gradivus—but cf. C.I.L. in loc), dedications to Venus, to Victoria Augusta (2584, 2585), to an unnamed deity by a sevir Augustalis and priest of Isis (2589) and to another deity unknown (possibly only the inscription of the side of a base was preserved) in 186 A.D., the name of Commodus being as usual erased (2590)—also the base of a statue with a dedicatory inscription to Tiberius set up in 30- 31 A.D., found by Biondi (2591), an epistyle with a dedication to the same emperor, erected two years later, by a curator lusus iuvenalis, (2592), a dedication to Nerva by the senate and people of Tusculum (2593), and fragmentary dedications to M. Aurelius, Septimius Severus, and Alexander Severus (2594, 2595, 2597) and some later emperor (2598)—the last found

1 Nibby, Schede, i. 64, saw it along the upper road, i.e. between the amphitheatre and the theatre, on October 13th, 1823 (?), and notes that there are traces of the stone having been cemented 362 THE BRITISH SCHOOL AT ROME. on the ancient road : further, the inscription of M'. Cordius Rufus, praetor and proconsul (he was, as the coins show, triumvir monetalis about 48 B.C.), who, at Tusculum, no doubt his native place, as is indicated by the representation of the Dioscuri on his coins, held the position of aedilis lustralis and monitor sacrorum (2603). In this connexion we may also note the discovery, before 1627, in antiquis Tusculi ruinis of an inscription erected by two aediles, recording an order given by them to remove (or destroy) false weights and measures {men- suras etpondera iniq{ud) {tolli or frangi) iusserunt—ib. 2625), and of a block of sperone found ' recently,' according to Mattei, i.e. early in the eighteenth century near the hill of Tusculum in the property of the Signori Meloni— not, I should imagine, the same as Monte Mellone below Monte Porzio {Papers, i. 253), bearing an inscription of a woman who was a sodalis iuvenum1 (2631—cf. 2635 and p. 247, supra), an inscription found in 1742 (but not apparently in the excavations of the villa above the Rufinella) recording the enlargement of a part of the Schola collegii dendrophorum by Sex. Octavius Felicianus, Senator and aedile of Tusculum and rex sacrorum, who had been made praefectus of this collegium (2634). Other inscriptions found by Biondi are—a dedication by resolution of the Senate made from the liberality of C. Fabius Passienus Saturninus, augur and aedilis lustralis {ib. 2628), a base of sperone, erected in 131 A.D. in honour of M. Pontius Felix, Senator, municipal aedile, sodalis {lusus iuvenalis) itemque aedilis et curator sodaliuni (2636) : 2637, a dedication to a magister aedituum Castoris et Pollucis, Augustalium honore functo was found earlier, no doubt by Lucien Bonaparte, as Nibby saw it in 1819 (it is now in the municipal collection at Frascati2) : 2642 is a fragmentary inscription, apparently referring to a statue, found in 1825, no doubt by Biondi. It is not my purpose to discuss in detail the large number of inscriptions certainly or probably found in the territory of Tusculum of which the exact provenance cannot be fixed. I may name a dedication to Castor and Pollux {C.I.L. xiv. 2576) made by Q. Fl(avius) Balbus, who was governor of Arabia {C.I.L. iii. 95) and consul suffectus at an unknown date,

1 The various lead tesserae relating to the Sodales Tusculani are collected by Rostowzew, Tesseraruni Plumbearum Sylloge, nos. 858 sqq. 2 The provenance of ib. 2639, a dedication to a curator (cf. 2629) by permission of the aeditui of Castor and Pollux, is not certain. Nibby gives a copy of it (Schede, i. 117) without indication of locality ; but from its subject-matter it may fairly be attributed to Tusculum. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 363 a list of 24 (?) men—exclusive of their praetores or praefecti, who con- tributed money for the repair of some building, in which the Augustales aeditui Castoris et Pollucis also joined, said to have been found at Ficulea {Papers, iii. 59 sgq.), but rightly, I think, attributed to Tusculum by Dessau (ib. 2620) ; an inscription recording the gift by one Caelia of a statue of her son Saufeius to the municipium to which she belonged, which coming from the Passionei collection {infra, 374) may or may not belong to Tusculum ; (2624) the sepulchral inscription of M\ Fictorius Synhistor cur{ator) aeditnont Castoris et Pollucis erected by the sixteen aeditui after his death, first seen in Rome in the seventeenth century, but, like 2620, attributed to Tusculum (2627). (On the other hand 2630, said by Fra Giocondo to have come from Tusculum, is with more reason attributed to Ostia: it mentions an official of the collegium fabrum tignariorum, of which we have frequent mention at Ostia, whereas there is no other inscription connected with it from Tusculum.) In April 1902 it was found that among the materials of which the altar in the chapel of S. Gregorio at Mondragone was built was a fragment of an inscription containing the fasti of a collegium as follows magiistri) Parhedrus glutin{ator) Demetrius topiar{ius) Appa topiar(ius) [M. Fur\io Camillo Sex(to) Nonio (Quinctiliano) co(n)s(iilibus) mag(Jstn) . . . Polit . . . The consuls mentioned are those of A.D. 8. The name of the collegium is uncertain : as the first of the three magistri was a bookbinder and the other two were gardeners, it contained a mixture of trades ; and Grossi-Gondi, who publishes the fragment, thinks that it may be part of the fasti of a collegium fabrum {Bull. Com. 1902, 99); but I see no reason for this.' Otherwise we have no mention of any collegia of members of a trade at Tusculum except the dendrophori, who were largely a religious associa- tion. This we should expect, as Tusculum was of course a residential and not a trading place. From these, and from our historical authorities, a short sketch of the history of Tusculum from the earliest times may be given, following in the footsteps of Dessau (C./.L. xiv. p. 252). According to tradition, the founder of Tusculum was Telegonus, the son of Ulysses and Circe The first mention of it in the Roman annals is at the end of the regal period, when it espoused the cause of Tarquinius Superbus on his expulsion from Rome, the chief of Tusculum, Octavius Mamilius, having received his daughter in marriage. Under Mamilius' leadership the thirty cities of the 364 THE BRITISH SCHOOL AT ROME.

Latin League waged war against Rome to secure the restoration of the Tarquins, but were defeated at the battle of the Lake Regillus {supra, 321) in 497 B.C. The history of the subsequent period is by no means clear ; Diodorus (xi. 40) tells us that in 484 B.C. Tusculum was taken by the Romans—an event not mentioned by Livy and Dionysius ; but we hear subsequently of close connexion between Rome and the people of Tusculum, except in 340 and 323 B.C., when they seem to have allied themselves with the enemies of Rome. It is clear that Tusculum was among the oldest of the municipia, if not the oldest: it already possessed the citizenship in 338 B.C., and according to Livy's account, with which other authors agree, received it as early as 381 B.C. When, however, it received the full citizenship is doubtful : after it did so, its citizens voted in the Papirian tribe, and seem to have been preponderant in it—though only one inscription from Tusculum or its neighbourhood mentions this tribe. Many of the chief families of Tusculum rose to the highest honours in Rome—the first was C. Fulvius, consul in 324 B.C.—and in time transferred their residence thither. Some of the cults of Tusculum were taken over by the Romans and placed under the charge of a college of Roman knights, who had no connexion with Tusculum itself. We hear little more of it under the Republic, except that in 211 B.C. Hannibal appeared before its gates, but was not admitted. At the end of the Republic, as Cicero says {Pro Plane. 8 § 19), Tusculum was full of men of consular rank ; but it had become so much a suburb of Rome that it had ceased to take any special pride in the honours which its citizens had won there. How favourite a summer resort it was, we learn from Cicero's works and letters. It may have come into vogue as such even before the Second Punic War: it is possible (but it depends on which version we adopt) that L. Caecilius Metellus, who was consul in 251 B.C. had an estate in the district of Tusculum,1 though whether it was a villa or a farm we do not know. (Veil. Max. i. 4 ; 4, 5.) During the Imperial period, as we have seen, the territory of Tusculum continued to be frequented even by the emperors themselves,2 though the town itself is hardly mentioned. The chief magistrates were aediles (we hear of a dictator only in the early annals), though the town council kept the name of Senate. The cult 1 One version of the text has Cum Metellus Tzisculanum peteret, the other Metello profuiscenti in agrum Tusculanum. 2 The description of Strabo (v. 3. 12, p. 239) may be noticed as a faithful picture. It is given in Tozer's Selections, p. 158. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—-III. 365 of Castor and Pollux had, as we have seen, a special importance, and their priests had also the charge of the worship of the imperial house. At the beginning of the third century A.D. we hear of a curator reipublicae Tusculanorum in the person of L. Marius Perpetuus. For the mediaeval history of Tusculum and its final abandonment in 1191 A.D. we may consult Tomassetti (Via Latina, 200 sqq.) and Seghetti (Tuscolo e Frascati, 105 sqq.). To attempt to determine the exact site of the discovery of the numerous brickstamps that are recorded as having been found ' at Tusculum ' would be almost impossible, and little would be gained by it. Among those discovered in December 1826 we may note C.I.L. xv. 265. 8, 359. 2, 1086. 4, 1088. 4, 1395- 3, for the reason that we have no other record of excavations then in progress ; while those shown by Rossini (tav. 71) are ib. 797, 822, 824. We may add the lamp ib. 6205 with the inscription annum novum faustum feliceni Ioviano. For sculptures seen at the Villa Rufmella by Wells cf. op. cit. 222. Among the many objects found at Tusculum, of which the provenance cannot be more exactly fixed, we may note the terracotta plaque in the Campana collection figured by Canina, Tuscolo, tav. iii. (the recognition of Ulysses) and the various terracottas given ibid. tav. l.-liii. { = EdifiziW. tav. xcvi.-xcviii.—one figure, xcvii., Fig. 8, occurring in Edifizi alone) : Hi. Fig. 3, which bears the stamp Vales {C.I.L. xv. 2553), is now at Aglie. There is, too, the statue of Dionysus with his left arm resting on a statue of Aphrodite, a larger replica of the Deepdene statue, now at St. Petersburg (Ermitage, no. 156, Clarac, 695, 1615 : see Michaelis, Ancient Marbles in Great Britain, p. 280, no. 3). Canina {Tuscolo, p. I45> tav. xxxv.) wrongly states that it is in the British Museum. Ficoroni {Bolla d'Oro, II, 12 = mem. 114 ed. Fea, Miscellanea, i. p. 172) mentions the discovery ' in the ruins of Tusculum ' of a ' vetro cimiteriale,' i.e. the bottom of a glass vase with two portraits in gold upon it, which was given to him by the Abbot Bernardo Sterbini. Among the proceeds of Lucien Bonaparte's excavations at Tusculum in 1818 (?) was a head of Antonina Augusta (?), sold by Capranesi to Mr. Rhodes, and seen in the possession of Col. Maitland Crichton by Michaelis {Ancient Marbles, p. 431 : Bernoulli, Rom. Ikonographie, ii. 1, p. 224, no. 16 merely repeats Michaelis' description). The bust of Drusus in the Capitol (Imperatori, 7) was also found here in 1818. 366 THE BRITISH SCHOOL AT ROME.

According to Melchiorri (Lettera al Sig. Cav. G. de Witte, intorno allo Stato del Mus. Cap., Rome, 1844, p. 10, no. 13), it belonged to a male heroic statue the rest of which is now in the porch of the Villa Borghese (no. xxviii. ?) It was purchased by the Conservatori and came into the museum in 1842. A statue found at Tusculum in 1820, and supposed to represent Cicero, has in reality no connexion with him. The head is figured in Kelsall's article in Classical Journal, xxiii. (1821) 265 ff. See Bernoulli, Rom. Ikon. i. 142. A double herm of Libya and a Triton, found 'near Tusculum' in 1829 and acquired by Bunsen from Capranesi, is also in Berlin {Beschreibung, no. 207). A painting ' from Tusculum,' seen at Naples by Helbig in the collection of one Barone, representing a youthful Bacchus and a panther, is described by Helbig in Bull. List. 1863, 150. The following sculptures from Tusculum out of the Campana collection are now in the Louvre—a draped Roman male figure (no. 920), a bust of M. Junius Brutus (no. 924 ' environs of Tusculum'), a bust of Otho (no. 1217), a bust of Caligula (no. 1227), a bust of Antonia (no. 1229). The Hellenistic bas-relief of Jupiter with the eagle at his feet from the excavations of Azara is also in the Louvre (no. 1365). A bronze tablet, said to have been found at Tusculum (to what it was affixed is uncertain), bore the inscription : Aincio (for Anicio) Probo v{iro) c(larissimo) et Anicihe (for Aniciae) Prob(a)e ciJarissimae) e [for f(eminaej\ with the monogram P F, the significance of which is still disputed, C.I.L. xiv. 4120, 2 = xv. 7157. The bust of Marius (no. 417 in the catalogue of 1880) in the Museo Torlonia is said to have been found at Tusculum. In the excavations made by Prince Aldobrandini (apparently not long before 1865) was found a bone tessera bearing the inscription Philetus Rutili sp(ectavit) k{alendis) Apiiilibus) Ti(berio) Plau(tio) (e)t Corivino) {consulibus') (45 A.D.), C.I.L. xiv. 4126. Among the inscriptions found at Tusculum is Kaibel, IG. xiv. 1120, a dedication

XVIII.—THE DISTRICT TO THE N. OF TUSCULUM AS FAR AS THE VIA LABICANA. To complete our survey of the territory to the N. of Tusculum, as far as the Via Labicana, we must examine (1) the road descending from the amphitheatre as far as the Barco di Borghese {supra, 326), (2) the road from the theatre to Le Cappellette, (3) the road running E. from these as far as Monte Porzio, and Monte Porzio itself. Taking the first of these, we first of all see on its right (N.E.)2 side some substructions in selce concrete, shown in our plan of Tusculum (31), and indicated also by Canina, who marks the road 'Via Particolare delle Ville ' ; then come some others on the right, above it, which form part of the side of a natural terrace on which there are no buildings : just below are the scanty remains of a large villa, by some mischance indicated neither by Grossi-Gondi nor by me : there is a large platform facing W.N.W., upon which a few remains of rooms are now to be seen, supported by a wall with curved niches, of opus reticulatum : a chamber behind this wall at one point is constructed of inferior brickwork. Below this is a small open reservoir, which lies a little to the S. of the track (marked in the map), which leaves the path ascending from the Villa Rufinella to the amphi- theatre close to the word ' Scuola,' and thence runs N.E. Below this track a group of ruins is marked in my map, which present some features of considerable interest. They are undoubtedly the 1 Stevenson (Vat. Lat. 10572, 11) saw the first of them at the Villa Rufinella in 1874. 2 The modern path passes through them, but the ancient road must have kept just below. 368 THE BRITISH SCHOOL AT ROME.

remains of a large villa, facing N.W. The N.W. front is supported by a wall 31'60 metres long,1 and yg$ metres high, of fine polygonal work of blocks of very hard red tufa, with close joints and the faces left a little rough, there being slight rustication in some cases. The largest blocks measure about roo by 0^46 metre, the smallest O'35XO'3i. The first 2 metres from the W. angle, however, are of small ashlar masonry of the same material: this angle is shown in Plate XXXII. Fig. 1. 2'OO metres before we reach the N.E. end of the wall there is a rectangular drain opening in it O'6o metre wide, and at least O'o,o high, roofed with a flat lintel, to which corresponds inside an arched passage lined with opus incertum. There is no wall on the N.E. side, as the ground slopes up towards it ; but on the S.W. there is a wall of opus reticulatum, and on the S.E. a cryptoporticus lined with the same material, to the S.E. of which is another wall in polygonal work, but less well preserved, originally about 18 metres in length and about i-5o in height as preserved, with opus reticulatum in front of it, forming, as Fonteanive tells us, the wall of a bathroom with a mosaic pavement. To the E., and a little higher up, are other rooms belonging to the villa and its entrance, with, perhaps, the pavingstones of a branch road coming down to it. Among the walls which are of brick and opus reticulatum of selce I found the brickstamp, C.I.L. xv. 911 b (first century A.D.). Here were found mosaics and other decorations (Tomassetti, 183 n.) and a brickstamp with Greek lettering {idem in Rom. Mitt. i. (1886), 63). In front of the wall on the N.W. there is a rectangular reservoir of selce concrete with external buttresses, three on the S.W., one at the W. angle, and five on the N.E. Grossi-Gondi speaks of it and marks it in his map as a ' tomb.' That the polygonal work is of Roman date and connected with the villa, I have not the smallest doubt (compare Papers, iii. passim) and Fonteanive {Avanzi Ciclopici, 86, who publishes a good photograph of the wall, which was cleared in 1886) and Giovenale {Diss. Pont. Ace. Arch. Ser. II. vol. vii. p. 332 and Fig. 5) are of the same opinion. Tomassetti, on the other hand, and Grossi-Gondi (p. 153) consider it to be of far older date, and the latter regards it as a portion of

1 Fonteanive gives the length at 23 metres : I suspect a misprint for 32. Tomassetti and Grossi-Gondi speak of one wall, 11 metres long and 3'85 high, having confused the two walls, apparently. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 369 the city wall of Tusculum. The former speaks vaguely of a line of walk of opus quadratutn which begins about 300 metres from the entrance gate of Camaldoli and after an interruption reappears in the part above the Villa Rufinella towards the amphitheatre, and proposes also to extend the city so far in this direction, but the idea seems to be quite impossible and unsupported by any evidence (cf. infra, 371). The material has been generally called sperone, but Giovenale calls it pietra albana (which should be peperino) or sperone in his text, and tufo vulcanico in the legend under his illustration. Below this villa the road must have turned N.E., though no actual traces of its course can be seen now (it is rightly marked in Grossi-Gondi's map, but not in mine), keeping just above the remains of another large villa. At the top of it is a modern house resting upon an ancient sub- terranean reservoir with three chambers. Below it is a large platform, and below that another, upon which are the remains certainly of one reservoir with three intercommunicating chambers and possibly, as Grossi-Gondi thinks (p. 154) (he marks the two lower reservoirs separately in his map), of another, the arcades of which, however, are now open, so that if there was ever a reservoir, its outer wall has gone. He made some excavations in 1898 {Bull. Coin. 1898, 337) and found various fragments of paintings and stucco from wall decorations and two brickstamps (C.I.L. xv. 2304, first century A.D.) and a fragment which is to be associated with ib. 1138 or 1869. Below this villa is a quarry in which a drain may be seen, running some io° N. of E. and belonging no doubt to this villa, and a small portion of the pavement of the road, running in the direction indicated by him. Thence it crossed the modern path from Mondragone to Camaldoli and descended to the valley to the N.E. of the former, which it followed in a N.W. direction, as Grossi-Gondi has made out from recent discoveries, and passing about at the point where are the letters ' nd' of Mondragone in my map, after which my marking is correct. Here are the remains of a building which he considers to be a tomb: then there is a villa in the right- angled bend of the modern road to Monte Porzio which he mentions (p. 155, n. 3) and marks in his map, but, I think, a little too far W.; and below the road to the N.W. is a reservoir in selce concrete, which we have both of us omitted to indicate, with three chambers originally (only two of

B B 37° THE BRITISH SCHOOL AT ROME. which are preserved), and nine arches in the dividing wall, running N.W. by W.: the vineyard is the Vigna Carocci, so that Grossi-Gondi is wrong in his marking and in p. 156, n. 1 (see supra, 329). The road then crossed the road to Colonna and continued towards Prata Porci {supra, 325). We have left unnoticed the Villa Mondragone, to which we must now return. Grossi-Gondi, who for many years was on the teaching staff of the Jesuit school established there, has written an excellent account of it1 and of the ancient remains upon which it rests, which he rightly attributes, from lead pipes found in 1732 (C.I.L. xv. 7847) to the brothers Quintilius Condianus and Quintilius Maximus, the owners of the great villa on the Via Appia (see my paper in Ausonia iv. (1910) 48 sqq.), who were put to death by Commodus about 183 A.D., whereas Stevenson (Cimitero di Zotico, 98) and others had thought that the villa at Barco di Borghese belonged to them. There are hardly any remains now visible,2 so that I may refer my readers to Grossi-Gondi's book, and to the supplementary notes which will be found in his Tuscolano, p. 142. Various antiquities have been found or preserved at Mondragone at different times: thus the head of a Muse in the Villa Borghese (Nibby, Mon. Scelti della Villa Borghese, Salone, no. 6, p. 42 : Venturi, Museo e Galleria Borghese, xxxvii) was brought thither from the Villa Mondragone, and the colossal head of Antinous in the Louvre (no. 1205) came from it through the Borghese collection. C.I.L. xiv. 2596 (a dedication to Caracalla by Aemilius Macer Faustinianus), now at the Villa Taverna, was first copied at Mondragone ; but Lanciani, Storia degli Scavi, iii. 49 suggests (ground- lessly, I think) that it may have been found on the Colle Cesarano with the inscriptions of the Caesonii {ib. 3900-3902 : cf. Papers iii. 138). He also enumerates other antiquities found or preserved there. To the W.N.W. is the villa marked Villa Borghese. It is also called Villa Taverna from its builder, Cardinal Ferdinando Taverna, who built it in 1604-5, but sold it in 1614 to Cardinal Scipione Borghese. It remained in the possession of the Borghese family until 1888, when it became the property of the Parisi family (Lanciani, Storia degli Scavi, iii. 53). It does not occupy an ancient site, and Grossi-Gondi 1 Le Ville Tuscolcme neW epoca classica e dopo il Rinascimento. La Villa dei Quintili e la Villa di Mondragone, Rome, 1901. His article in Bull. Com. 1898, 313 sqq. is repeated in this work (pp. 20 sqq. and 285 sqq.). 2 There are scanty traces in the path from Camaldoli on the N.E. side of the garden. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 371

(p. 143) is probably right in supposing that the land round it belonged to the Villa of the Quintilii. A number of inscriptions and other antiquities are, or have been preserved there, among which I may mention the sarcophagus with a representation of the bringing home of the body of Meleager seen there by Robert in 1882 {Sark.-Rel. iii. 299) but since removed, the Lycurgus sarcophagus (Matz-Duhn, 2269), and others men- tioned by Tomassetti ( Via Latina, 178). To the N.E of the house is a large reservoir: Lanciani {Bull. Com. 1884, 186) and Grossi-Gondi {ib. 1898, 334; Villa dei Quintilii, 289) conjectured that it supplied the villa at Barco di Borghese. A plan of it is given by Canina (Tuscolo, tav. ix. = Ediftzi, VI. tav. lxxxi.), but Grossi-Gondi does not indicate it in his map (for its site cf. Canina, tav. vii.). At the two points between the Villa Borghese and the Barco where he marks ruins, there are traces of platforms with debris, but no remains of construction are visible. For the so-called Villa Vecchia cf. Lanciani, Storia degli Scavi, iii. 48.1 With the Barco di Borghese we have already dealt {supra, 327 sqq.) and we may now return to the theatre of Tusculum, and follow the road 2 which descended from it past Camaldoli, the beginning of which has been described supra, 356. We may begin our consideration of it with the scanty ruins marked 9 in our plan, which are situated immediately to the S. of the enclosure wall of the monastery of Camaldoli (cf. supra, 350) and probably belonged to the concrete of a villa platform, though nothing but debris is now visible there. Just above them I saw in 1898 (they are not now to be seen) some scanty remains of walling of blocks of tufa about o-6o metre (2 Roman feet) thick ; and there were similar remains further N.N.W. These are, I imagine, the traces of the supposed city walls of Tusculum to which Tomassetti alludes, but, as I have said, I cannot agree with him that this is what they represent. To the W. is the so-called Villa of the Caecilii (7 on our plan : supra, 348), to the N.W. of which (omitted in the map) and just below the tomb marked 8 3 is a building in opus quadratum and opus reticulatum—no doubt

1 Tomassetti gives {Bull. Com. 1892, 359) a fragment of a Greek inscription, the significance of which is not clear, built into the casino of the Villa Vecchia, on the road from Frascati to Monte Porzio. 2 C.I.L. xiv. 2693 (a fragmentary sepulchral inscription found by Biondi in 1830) must belong to this road. 3 The tomb 8 is marked too far W. in our plan, but this is due to the fact that there are errors B B 2 372 THE BRITISH SCHOOL AT ROME. a villa on the S. edge of the modern track from the Villa Rufinella {supra, 367). To the N.E. (it is shown on the W. edge of Map I. of the present volume) is a large villa in opus reticulatum, marked Villa dei Furii by Canina {Edifizi, VI. tav. lxxx., lxxxv.; but not in Tuscolo, tav. vi., vii., which was published in 1841 1: see below). The ancient road passed within the enclosure wall of the wood belonging to the monastery of Camaldoli, and Canina {Tuscolo, p. 134) alludes to it. Further remains of it were found in 1901 (Grossi-Gondi, p. 160) at a considerable depth, varying in width from 2 to 4 metres.2 In the upper part of its course it made a sharp turn to avoid an ancient villa not marked in my map, but indicated by Grossi- Gondi, and described by him (p. 166, cf. Bull. Com. 1902, 103). This building was discovered in 1901, and its plan was entirely different from the normal plan of a Roman house : among the rooms were two halls, paved with mosaic, with geometric designs in black and white. The brickstamps which were found in the building belonged to the first century A.D. ; but some blocks of sperone which were found there may, Grossi-Gondi thinks, have belonged to some earlier building. To the S.E. of it was a large cistern measuring 27-60 by 5'6o metres ; and Mattei (Tuscolo, 78) describes a system of rock-hewn cisterns found by the monks in 1704 in the wood, consisting of passages j\ palms (1*65 metre) wide, in which a man could walk, with shafts at intervals. Chaupy {op. cit. ii. 206) also mentions them. It seems to me that the chamber discovered in 1862, containing a pavement in black and white mosaic, and representing gladiatorial scenes, must have belonged to the same building: it is said to have been found at the S.W. extremity of the enclosure of Camaldoli, towards the so-called Acqua del Cardinale. The pavement is now in the Villa Aldobrandini (Grossi-Gondi, pp. 19, 165 and tav. iii.: cf. Ann. Inst. 1863, 397). The question as to the site of the tomb of the Furii is fully discussed by Grossi-Gondi (161 sqq.~). The tomb had a vestibule of opus quadratum, with a carefully closed door of peperino, and was itself a chamber hewn in the tufa, in which was a sarcophagus some 5 feet long, and several smaller urns, about 12 in number : on some of these were the inscriptions C.I.L. xiv. in the plan of Canina from which it is taken, which could not be corrected without a more extended survey than I was able to undertake. 1 There he marks the tomb of the Furii just inside the wood of Camaldoli, at its S.E. extremity. 2 Stevenson [Vat. Lat. 10572, 9/) records its discovery in January 1894. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 373

2700-2707, 2750. The discovery was made in 1665 or 1667 (?) (Kircher, Vetus Latium, 671; Falconieri, Inscriptiones athleticae in Gronovius, Thesaurus Ant. Graec. viii. 2348) in the wood belonging to the monastery, close to the ancient road ; but the wood was in those days apparently larger than in the nineteenth century, the present enclosure wall not including the site of the discovery. For in 18422 Canina, who was excavating on behalf of the Queen of Sardinia {supra, 359)—and not primarily of Campana, as Grossi-Gondi supposes—found outside it, at the point where he marks 'Villa dei Furii,' two columns with dedications in identical terms to Fortune and to Mars (M. Fourio{s) Cf. tribunos inilitare de praidad dedet, C.I.L. xiv. 2577, 2578), and with them some fragments of statuary, including a fine standing statue of Jupiter, which was conveyed to Aglie, and a fragment of the base of a statue with the name of the artist Sosicles (Loewy, Inschr. Gr. Bildli. no. 434). It was on this occasion recognized that the site of the discovery belonged to the monastery, and it was actually restored to it {Edifizi, V. p. 72, n. 13). The discovery of the dedications renders it probable, though not certain, that either the villa indicated by Canina, or that described by Grossi-Gondi, belonged to members of the gens Furia, though the inscriptions belong to the first century B.C., while the remains of villas point, as the latter recognizes, to a date about a century later. The line of the road probably cut that of the avenue leading to Camaldoli, inasmuch as Canina, in the map cited {Tuscolo, tav. vii.) marks a tomb a little below it, but in the same line. Beyond that point, we may conjecture that it descended towards Le Cappellette, as Grossi-Gondi does, or simply suppose that it ran towards Mondragone, as I have done (cf., however, supra, 326),3 Nibby yAnalisi, iii. 295, 597 ; Schede, i. 63) considers the direct path from Camaldoli to Frascati, i.e. that which passes S. of the Villa Falconieri, to be ancient ; but this seems to me uncertain : there are some pavingstones in it to the S.E. of Mondragone, but they are not, apparently, in situ. The monastery of Camaldoli was originally erected in 1611 : a part of 1 Cf. also the additional information from his MS. notes published by Fea, Miscellanea, i. 321. 2 1852, the date given by Canina himself, is the result of a misprint. 3 It is not improbable that Holste is referring to this road {Cod. Dresd. F. 193 f. 67 ' da Grottaferrata andando sotto Frascati verso l'osteria del Finocchio si vedono per tutto strade antiche, prima e la Latina . . . poi quella del Tuscolo che passa accanto a Frascati poi un' altra che tira su per la valle verso Monte Dragoni dove si vedono vestigi immensi di una villa grandissima' (Le Cappellette). 374 THE BRITISH SCHOOL AT ROME. it was used as his villa by Cardinal Passionei, who collected there various antiquities, including many inscriptions, about 800 in all, and a library, etc. (Tomassetti, 186 n., Lanciani, Wanderings in the Roman Campagna, 294). Most of the inscriptions are now in the Vatican. After the Cardinal's death on July 5th, 1761, his antiquities were at once put up for sale : John Adam in his journal (published in Diary of Fine Arts ii. nos. 9, 10 (1831)), mentions that the sale was taking place on the 24th of that month and on August 27th says ' we made another jaunt to Frascati, to look at some of the antiquities of Cardinal Passionei, and made offer of 205 crowns for the vase of the Villa Adriana and some other fragments.' Ghezzi notes the discovery in making an aqueduct from Tusculum to Camaldoli in 1741 of a Greek middle bronze of Trajan ' questa medaglia fu trovata nel condotto che fa fare il Sigr. Cardinal Passionei p(er) far con- durre l'acqua la quale viene dal Tuscolo p(er) condottarla e p(er) farla venire al suo romitorio nei Camaldoli di Frascati, il di' 29 Xbre 1741.' {Cod. Brit. Mus. f. 2. published by Lanciani, Bull. Com. 1893, 169.) Ghezzi also noted and drew {Cod. cit. f. 20) a marble mask used as a fountain jet 'questa testa . . . e posseduta dall' Emmo. Sr. Cardie. Passionei, il quale l'(h)a collocata in una fontana che getta l'acqua nel sarcofago che gli dono N. Sre. P. Benedetto XIIII, la quale stava in faccia alia pigna al belvedere al Vaticano et il med° . . l'(h)a fatta collocare al fine della macchia nel suo romitorio di Frascati, in faccia al Pischierone copiosa di diverse sorte di pesci, e da belvedere di San Pietro per condurla alii Camaldoli gl' importo' per la sola conduttura scudi 30 et io cav. Ghezzi me ne sono lassata la memoria, essendo una testa non sola bella, ma assai particolare.' Other antiquities in Cardinal Passionei's possession, but from Rome, are noted ibid. 21, 104, 111-118, 121. On f. 120 he gives a drawing of a fragment of a marble statue of a boy with an elaborate sandal, leaning against a tree trunk—' questo frag- mento lo possiede Fra Bonifacio laico nel romitorio dei Camaldoli di Frascati, al quale gli fu dato da un villano che disse haverlo cavato al Tuscolo . . . il quale fragmento e bellissimo e particolare p(er) il sandalo: il medm° e di marmo Pario alto 1 palmo e un quarto . . . (18 ottobre, 1741).' The Greek inscriptions collected by Passionei at Camaldoli include Kaibel, I. G. xiv. 1360 ('AicpiSt, Kvvaplm—the tombstone of a little dog—seen in 1858 by Vischer at Ockham Park, Surrey, the seat of Lord Lovelace), CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 375

1544, 1802 (unimportant tomb inscriptions, now lost), 1982, 2026, 2031 (now in the Galleria Lapidaria of the Vatican). An interesting Christian sarcophagus found at Camaldoli and trans- ported to the Villa Taverna is published and described by De Rossi {Bull. Crist, 1872, pp. 133, 140 and pi. vi. of the French translation). Stevenson {Vat. Lat. 10572, 141) notes a report of Pasquale Antim (dated September 8th 1889) that the monks of Camaldoli had found a male statue of a wrestler (?) and other antiquities ; and here Helbig copied in 1888 the early brickstamp (beginning of first century A.D. or earlier) C.I.L. xv. 824. 7. The road from Frascati to Monte Porzio has been already dealt with as far as II Barco di Borghese {supra, 327). From this point, as Grossi- Gondi notes (p. 145), the line of the modern road is entirely different from the ancient, which at first followed the line of the lane on the E. of Mon- dragone ; along this, in the Vigna Mancini, some pavement was found in 1888.1 Beyond this point, however, its course is quite uncertain and both Grossi-Gondi and I {Papers i. Map V.) show it conjecturally : we can only be sure that it did not follow the modern line, inasmuch as this last cuts through an ancient villa just to the N. of the villa of the Collegio Nazionale {infra, 377), while the older track kept above the great villa which we now reach, the remains of which bear the name of Le Cappellette (Raggi, Sui colli Albani e Tuscolani, no). It consists of two large platforms: the lower of them, to the S.W., is supported on the N.W. side by a wall with large semicircular niches, faced with good opus reticulatum of selce, with quoins of the same material, and bands of brick : the arches are of fine brickwork. The upper platform, to the N.E., is supported on the N.W. and S-W. sides by lofty arches and pillars of opus reticulatum of selce with selce quoins and voussoirs, which is possibly earlier than the work in the lower platform : these have been reinforced later within with concrete with the same facing, but inferior in quality, and with the addition of bands of baked bricks: on the N.W. side these additions are in two stories. A view of the N.W. side is given in Plate XXXII. Fig. 2. In 1888 seven fragments of lead piping were found in the Vigna Mancini, which occupies this terrace, bearing the inscription Matidiae Aug. fil. Ixxxii. . . . Salon{ius) Epictetus f{ecit) {C.I.L. xv. 7822): whether the

1 Not. Scavi, 1888, 141 speaks only of ' poligoni di peperino,' a material never used for paving- tones, but Grossi-Gondi no doubt obtained better information on the spot. 376 THE BRITISH SCHOOL AT ROME. reference is to Matidia the elder or the younger, i.e. to the niece or the grandniece of Trajan, is uncertain; but we thus have the name of an owner of the villa. Remains of painted decoration were found at the same time. For discoveries of statues etc., made here in the seventeenth century (the objects are said to have been transported to the Villa Borghese), see Cod. Tusc. tit., quoted by Lanciani, Bull. Com. 1884, 211 sqq. Kircher ( Vetus Latium, 73) had attributed the ruins to the temple of Bona Fortuna mentioned by Dionysius of Halicarnassus, and the theory is repeated by Pococke {Brit. Mus. Add. MS. 22981, 55): it is, however, without founda- tion, and Nibby (Analisi ii. 356) clearly saw the real nature of these remains. From the Atti del Catnerlengato, Tit. iv. fasc. 1668 we learn that in March 1832 Don Camillo Borghese asked for permission to excavate in an oliveyard belonging to him in the territory of Monte Porzio, Contrada Le Cappellette, and more exactly in Contrada Pentima de' Frati, which lay on the old road to Monte Porzio. Rossignani reported that although there were large ruins close by, with double arches one above the other, no harm could be done to them by excavation. He added reports, that Campana (the tenant) had opposed, or that Campana himself had dug by night, or that Campana and the Prince had agreed, and that really the former was working with Pietro Paolo Spagna. In any case leave was granted ; but we have no report of the result. Grossi-Gondi attributes to the reservoir of this villa a large wall which he has noticed in the vineyard above (p. 189). To the S.E. of this villa are the remains of another large platform (marked on the extreme E. edge of Map II. of Papers iv.).1 The Casale Campitelli rests on an ancient reservoir (marked R in Papers i. Map V.). To the E. again, beyond the Casale Campitelli, are the remains of a platform of polygonal work, belonging to a villa marked in Papers i. Map V., to which I must refer ray readers for the rest of the territory to be described in the present section. Excavations were made here in 1840 (Canina, Tuscolo, 107, 108 cf. tav. vii.) and in the Atti del Camerlengato fasc. 1668 is a letter from Canina dated February 3rd, 1841, stating that Prince Borghese was continuing with a few men the work which he had

1 The Casino del Collegio Clementino which Nibby saw high above him (Viaggio, ii. 58, ought to be, I think, the Casale Campitelli. Chaupy speaks of a Villa Vallemani here or hereabouts (ii. 218). CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 377 begun in the previous October on the hill belonging to him above the village of Monte Porzio towards Camaldoli, which was in the territory of Monte Porzio: that a large villa had been discovered which lay on the road which ran from the Via Labicana to Tusculum,1 and that several marble pavements of various colours, and fairly well preserved, had been found. To the N. again, and to the E. of Le Cappellette, is the villa of the Collegio Nazionale (formerly the Villa Lucidi) built upon the platform of a large ancient villa, which has arcades of opus reticulatum, and is cut through by the modern road. Grossi-Gondi (p. 189) refers to the ancient road which perhaps passed S. of the site, the sepulchral inscription C.I.L. xiv. 2757 {T, Vinicio Corintho optimo et karissimo lib(erto) patronus which was found to the S. of the Villa Lucidi in 1879), and considers that it may give some indication of the ownership of the villa—but, I think, with no more foundation than in other similar cases. We find from Atti. cit. fasc. 2362, that in July 1835 permission was granted to Don F. Borghese Aldobrandini to excavate in a vineyard and oliveyard of his in the territory of Monte Porzio occupied by Campana ; but again we have no record of the results except that given by Canina, Tusculo, 104, who tells us that the site was a little nearer than Le Cappellette to Monte Porzio itself, that a few worked marbles were found, and that at the end of the previous century Campana had found pavements of mosaic and of fine marbles there. The site was obviously not the Vigna Lucidi, and it may have been at point 351, N. of the road. Here are the remains of a villa (Grossi-Gondi indeed marks two groups of ruins close together in his map), in which I saw the gutter blocks and columns in peperino, of the peristyle, and the torso of a nude male statue (probably sepulchral), of which the front alone was worked. In 1826 the breaking up of the ground in the Vigna Lucidi resulted in the discovery of some statues—a Leda and the Swan (Canina, Tusculo, p. 104, 146, and tav. xxxv.—to whom the provenance is due), a torso of Apollo, and a colossal torso of a seated male figure (an emperor ?), described by Nibby, Monumenti Scelti della Villa Borghese, Portico no. 6, p. 16, Camera iv. no. 5, p. 112 : cf. Reinach, Repertoire ii. 416, 3 : Venturi, // museo e la

1 This road is marked by Canina, but I know of no further evidence for its actual existence (infra, 405 : cf. Tuscolo, p. 68). 378 THE BRITISH SCHOOL AT ROME.

Galleria Borghese, xv., xxvii. (the latter also states that ii., a fragment of another statue of Apollo, and xvi., a nude torso, were found at the same time : with regard to the former Nibby gives no indication, while with regard to the latter, which is probably the second of the three torsi mentioned by him under no. 6, Venturi has probably misinterpreted Nibby, who expressly distinguishes it from the other two). Excavations were then made systematically under Canina's charge, but nothing of importance seems to have been found. From A Hi del Camerlengato Tit. iv. fasc. 3422 we learn that in April 1845, it came to the knowledge of Don Marcantonio Borghese that the Padri Somaschi del Clementino, tenants of the Villa Lucidi, had excavated without permission and found various objects of art. The custode and vignarolo of the villa stated that in breaking up the ground anew in the vineyard, two animals of marble, lacking legs and portions of the head, were found—also the pedestal of a column, a marble slab broken at the corners, various fragments of marble, brick, and three large tiles, which covered the opening of some grottos. The Governor of Frascati examined the question, and sequestrated everything except a capital of a column, already conveyed to the Collegio Clementino, but reported that the objects were of no importance. Further E. again, to the S. of Monte Porzio, is the former Villa Gammarelli, now a convent: just above it, to the S., within its garden, is a platform of selce concrete facing apparently N.W. by W. ; on the N.E. by N. side a very fine piece of polygonal work in selce is preserved, which is undoubtedly of Roman date and formed part of the facing : the jointing is very good, the angle of one block being frequently let into the next block, and the fronts of the blocks are bossed : some measure as much as O'9O by o-6o metre, but 070 by o-5o is an average dimension. The walls of the villa above were faced with opus reticulatum and opus incertum. To the S. again, close to the boundary of the property, recent excavations have brought to light the remains of some underground chambers, possibly connected with a water supply, but not completely cleared and now filled up again. A space measuring about 7 metres long (there being no wall on the N.W., the width could not be determined) is lined with walls of concrete on the S.W. (short side), with brick on the S.E. (long side), and with opus reticulatum on the N.E. (short side). All the walls go CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 379 down straight, except that on the N.E., in which, 3'6o below ground level, is an arched opening i-55 metre in span and 192 metre in depth, and 3-80 in height, with a horizontal strut of masonry across the opening, i-4O below the top of the vault. This goes in some 2 metres further and is then filled with earth, being lined with cement as far as it can be followed. It had a branch to the N.W., which has not been cleared. Under the floor of this chamber is a small drain about 020 metre in diameter, and the excavation (in search of a water supply) had been carried some 4 metres deeper. Above, there was a cement lined space O'6o metre wide on the N.E. side, into which a water channel, 0^48 wide and O'6o high, discharged ; while on the S.W. there was a channel 1 metre wide, beyond which was found a late burial, the body being placed under tiles (one of which bore the stamp C.I.L. xv. 2321), and a minim of the 4th century A.D. having been found with it. Just outside the boundary to the S.W. of the villa, on the N. edge of the chestnut wood on the slopes of the Monte Ciufoli, is a small single- chambered reservoir of selce concrete, with buttresses on the lower side. On the E. of the villa, near the 'G' of Gammarelli, a path (not marked in the map) passes by and over the debris of another building, no doubt a villa. To the N. of the Villa Gammarelli, under the house at the'point where the road to Monte Compatri leaves the road to the village of Monte Porzio, are the remains of a reservoir. The village occupies a conical hill, which might well represent an ancient site, though we have no reason for supposing that it does so. A photograph of it from the S.W., which I owe to the kindness of Miss Dora E. Bulwer, is given in Plate XXXIII. Fig. 1. The name of Monte Porzio appears, perhaps in the Register of Gregory II., and certainly about the middle of the eleventh century (Tomassetti, Via Latina, 261), in the form Mons Porculi, and the addition Catone is modern, being due to the desire of antiquaries to place on this site a villa of Cato. In the village there are no ancient remains in situ, as far as I know. The origin of the name is no doubt the same as that of Prata Porci {Papers, i. 244), and the derivation from Porcius is not at all improbable. Cicero (De Rep. i. 1) says M. Catoni . . . limit Tusculi se in otio delectari salubri et propinquo loco, which gives us no precise indication of the site of his house, and hardly, I think, justifies Grossi-Gondi (p. 187) 380 THE BRITISH SCHOOL AT ROME. in supposing that Cato rather possessed a house within the town of Tusculum itself. The antiquities to be seen within the village of Monte Porzio are noted by Tomassetti (260 n.). There are also various antiquities which are recorded as having been discovered or seen at or in the territory of Monte Porzio, without a more accurate determination of their provenance being possible. We may enumerate a torso of Hercules, seen in the smaller Palazzo Mattei by Matz-Duhn (no. 104): some mosaic pavements men- tioned by Furietti {De Musivis, Rome, 1752, p. 58) as existing in the plain at the foot of Monte Porzio : the brickstamp C.I.L. xv. 1250 (first century A.D.), 2336. 3 (end of first century). We may add two statues of Venus, one in the Vatican, the other in the Louvre, which, according to Marocco {Stato Pontificio vii. 143), were found in the territory, but which I have not yet identified. To Monte Porzio, though their exact provenance is uncertain, belong the inscriptions C.I.L. xiv. 2580 (a dedication to Mars Gradivus by L. Plutius Pius . .. aedilis, monitor, augur, praef{ectus) sacrorium)—no doubt of Tusculum, cf. ibid. 2603), 2587 (a dedication to some deity not mentioned). The statue formerly supposed to represent Sextus Pompeius, which came to the Louvre (no. 1251) from the Borghese collection, was found in the territory of Monte Porzio apparently at the end of the eighteenth century: it bears the inscription of the artist, Ophelion son of Aristonides. (Bernoulli, Rom. Ikon. i. 226, Loewy, I.G.B. p. 289. no. 432, Kaibel, I.G. xiv. 1277.) Under the village of Monte Porzio on the S.W. are inconsiderable ruins (buttresses in concrete) marked in the map. To the W. is the house marked Tavolacci, to the N. of which are the remains of a building on the point of the hill. To the N. of the village, under the Casale Albertazzi, are considerable remains of a villa, including a large long reservoir, divided into two chambers each 1 "70 metre wide, by pillars which carry the vaulting, of which Steven- son {Vat. Lat. 10572, IS7T) gives a sketch plan and some particulars: cf. Tomassetti, loc. cit. To the E. of the village, on the summit of the hill called I Romoli, 391 metres above sea-level, are the remains of another villa, in opus reticu- latum, now less well preserved than when Stevenson saw them {cod. cit. f. 159, 160): the terrace occupied a space measuring some 100 by 60 metres, and in the centre of it is a reservoir 13 metres long, with two aisles, 2'8s CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 381 and 2^63 metres wide, of which he gives a plan. To the N.W. (near the 'a' of Catone), he noticed that the breaking up of the ground for a new vine- yard had led to the discovery of the debris of yet another building; and to the N. of point 39] is a possibly ancient quarry indicated in the map, and described by Stevenson {Vat. Lat. 10563, 23). S.E. of point 391, N.W. of the Casale Palocci, is the debris of a villa ; while to the N. of it, on the S. slope of Monte Doddo, is the concrete of a villa platform, facing due W., with a drainage passage in it, 1^40 metre high and C62 wide. I may note that Stevenson (Vat. Lat. 10563, 23) carefully examined the path from I Pallotta (or Pal- locci, as he calls it) to Monte Compatri (the old road to Monte Porzio) and found traces of ancient paving. The district to the E. of this point has already been described in Papers i. 259 sqq., but I may perhaps give a few additional notes upon it in this connection, largely based upon a further examination in which I followed out the hints given in Stevenson's notes and maps. I subjoin a plan of the reservoir under the Casale Brandolini: to the S. of it are ruins in opus reticulatum of the villa which it supplied : among these were found later burials ; one of the bodies was apparently covered by a much worn slab of marble, the inscription upon which I copied as well as I could.

006 D M

ASSI.AEHERO 004

0-076 CASSIO-POI OIV&I-DVLCIS TCVRATom ASSIAE-I Ml!!!AE El SSIMAE RTIS ERISQ RVM N. of the villa marked Le Cappellette in the map there is a large open reservoir. 382 THE BRITISH SCHOOL AT ROME.

To the S., N. of the Fontana Laura, are mediaeval ruins* and further N.E. again, to the N. of the Colle di S. Andrea, Stevenson notes the existence, in the vineyards and in the fieldwalls, of debris of buildings and of pavingstones—the latter probably from the Via Labicana. On the summit of the Colle di S. Andrea itself are the remains of a villa with a cryptoporticus. Taking the path from this point we find near the ' P ' of ' Piano della Faveta,' the remains of a villa noted by Stevenson, and further S. what may be the edging of an ancient road going E.S.E. across the path ; while near point 366 other ruins have recently been destroyed. Further up, on the E. edge of the Colle Tufino, is the actual pavement of an ancient road ascending southwards, the course of which cannot be ^^^^^^^^^^^^^ followed very far. On the top of the hill there 1 has been an extensive villa : on the E. edge, """""""""I W. of the line of the road, is a black mosaic 1 f , " floor, and under the ruined house on the summit '••" "''" ~ "•"" is a large reservoir in selce concrete, of which I give a plan in the text. It runs N. by W., the W. side being built against the hill. Further W. is the hill marked I Colli di Fontana Molara {Papers i. 264): on the top of the hill are two narrow underground passages (no doubt for the storage of water) lined with cement. The wood to the S.W. is known as the Macchia Pietreliscie, which should point to there having once been an ancient road found in it, but of this I could learn nothing.2 There is the debris of a building in it, near the point of one of the zigzags of the modern road ; and a little to the N. of the Casa Lutti (see Map I. of the present volume) there is a vaulted substruction of selce concrete going N. and S., with a ruined house above it. Finally, to the N. of the Convent of S. Silvestro3 {Papers i. 263) below a high cliff, and above the high-road, are the remains of what may be a cryptoporticus with two aisles : my guide, however, gave it the name of

1 Stevenson (Vat. Lat. 10572, 164) gives from a copy by Pasquale Antini, the following inscription as found at La Caricara (Papers i. 262) : DlS M VILESIA MATERI PIISSIME FECIT, etc. 2 Can this be the road alluded to by Chaupy (Maison de Campagne a"Horace, ii. 184) as running between Monte Compatri and Rocca Priora (infra, 411 n.)? 3 There are two reservoirs, not one only, to the S.S.E. of it, on the way up from the village, and some substructions also. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 383

S. Francesco, it being supposed traditionally to have formed part of a long ruined monastery,1 and Stevenson ( Vat. Lat. 10563, 23) who saw it on Aug. 27, 1890, considered it mediaeval, and very likely rightly. We have still to deal with the eastern portion of the road from Frascati to Colonna, as far as it falls within Map V. of Papers i., and with the ancient remains along it.2 (As to its antiquity see Papers i. 253, n. 1 and supra, 326.) The Casale Celli occupies the site of an extensive ancient villa, and remains of black and white mosaic pavement may be seen there: to the N.E. is a large round reservoir, at the ground level, on a terrace supported by a wall with niches of opus incertum. Various architectural fragments (columns, capitals, etc.) were found, and seen and described by Stevenson ( Vat. Lat. 10572, 37 sqq.), at the time of whose visit it belonged to Leonori, having previously been the property of one Anderson,2 and in 1777 of the brothers Gizzarelli. In the doorway into the yard to the S. of the house both he and I copied a fragmentary inscription, which I read thus upon a piece of marble C15 by o-o6 metre, the letters being O"O3 high. A number of large lead pipes were found, with an internal diameter of o-io to 0*13 metre, which were said to have come from the direction of the Casale Albertazzi {supra, 380). Further E., at the Casale Janari, also to the N. of the modern road, there are only a few marble fragments visible, and it is doubtful whether the site is an ancient one ; but the Casale Brandi (formerly Casale Filonardi or Lionardi or Tofanelli) is built upon ancient ruins in opus reticulatum, and under it are extensive reservoirs formed of several small chambers of roughly faced selce concrete. Built into the modern wall on its W. side is the marble cippus with the sepulchral inscription C.I.L. xiv. 2616, erected to P. Avidius Trachalus, praefect of a cohort and tribune of a cohort, and

1 The reader need hardly be reminded that the current distinction in English, according to which monastery is used for a community of men, and convent for a community of women (whereas Italian usage is rather in favour of the reverse), has no logical nor historical foundation. 2 I do not know the site of the Vigna Bellini (though perhaps it fell within the territory of Frascati) in which in 1879, in making the road from Frascati to Colonna, there were found (and removed to the Museo Kircheriano in Rome) 17 fragments of marble decoration, parts of two statues, a spur, and a lead pipe {Not. Siavi, 1879, 206). 3 Tomassetti (p. 260, n. no. 9) tells us that an inscription of one Pomponius had recently been found in this vineyard, but lost again. 384 THE BRITISH SCHOOL AT ROME. two other inscriptions from tombs {id. 2719, 2738), one on a cippus, the other on a sarcophagus, have been seen there. The fountain basin is a plain sarcophagus, measuring 2-40 by roo. metres over all, with spaces for two bodies. Stevenson {cod. cit. 39) copied here in front of the casale the inscription : SER-SVLPICI HOMILI which apparently was carved upon the left half of a double tablet, the right half of which was without inscription. He also noted here various architectural fragments. A modern path leads hence northwards to the Casale Grandi, which lies E. of the Casale di Fontana Candida. There is no proof of the antiquity of either of the paths to Fontana Candida: W. of the westernmost is the Casale of La Quercetta which certainly occupies the site of an ancient building. To the S. of the Casale di Fontana Candida is a mediaeval ruin, on an ancient site, and to the E. of this some remains of opus reticulatum walls. Just to the N. of the casale the railway curs through a large villa of the same material, which extended under the casale itself. In the courtyard of the casale I saw a draped male toga statue and some fine terracotta antefixae.1 In 1883 a marble slab was found at Fontana Candida, bearing the following inscription : D{is) M{anibus) Crescenti Sili Italici Collegium salu- tarem {sic). {C.I.L. xiv. 2653.) The Silius Italicus mentioned is either the poet, who was consul in 68 A.D., or his eldest son, who was consul suffectus before 101 A.D. (see Prosopographia iii. p. 245, no. 509). Lanciani (Wanderings in the Roman Campagna, 266) wishes to identify the collegium salutare of this inscription with the well-known collegium salutare cultorum Dianae et Antinoi {C.I.L. xiv. 2112), in the statutes of which there is a provision that if a member died within twenty miles of Lanuvium, his body should be brought to the headquarters of the collegium and buried with the usual forms. This is, of course, possible ; but as De Rossi had already pointed out {Bull. Com. 1882, 144), it is by no means a necessary supposition, inasmuch as we have records of the existence of other collegia salutaria—Waltzing indeed {Corporations

1 Prehistoric tombs of the iron age have been found in the Vigna Giammaria in the district of Fontana Candida : the vases are preserved in the archaeological museum at Frascati, and are unpublished. {Not. Scavi, 1902, p. 171, and note 3.) CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 385

Professionnelles, iv. 202) enumerates 19, all of them in Italy but one in Gaul, at Narbonne, and another in Portugal, at Coimbra. It is generally admitted that the epithet salutare was chosen for the sake of a good omen and signified that the object of the collegium was to guarantee a decent burial to the members. With this inscription were found the brickstamps C.I.L. xv. 1443 (end of first or beginning of second century), 2260 (first century A.D.), according to Lanciani, who copied them in the shop of Alessandro Fausti, a dealer in antiquities at Frascati {Not. Scavi, 1883, 85). They are certainly not later in date than the inscription, the lettering of which seems to point to the second century (though it cannot be very late in it), and it is a pity that we have not further information as to the circumstances of the discovery. In a vineyard near Fontana Candida was found the unimportant sepulchral inscription C.I.L. xiv. 2680, which is still preserved there; while another (id. 2785) was removed from Fontana Candida to Frascati in 1884, according to Not. Scavi, 1884, 348 (a statement slightly inconsistent with Dessau's account, C.I.L. in loc). Beyond Fontana Candida (below Monte Porzio) in the property of one G. B. Corvaccini, according to Grandi, the unimportant sepulchral inscrip- tions C.I.L. xiv. 2748, 2758 were found about the year 1846 {supra, 319); and below it in the vigna formerly Vivari, now Montani, are the remains of a villa (not now so conspicuous as Grandi makes out—see Papers, i. 252 and Map V.), near which was found, according to Grandi, in 1840, the unimportant sepulchral inscription C.I.L. xiv. 2695. Another inscription of this nature from Fontana Calida (sic) is given in Not. Scavi, 1895, 249. Both Stevenson (in 1875—Vat. Lat. 10572, I6IV) and I have seen and copied at the Casale Grandi a little further E., the inscription : IN FR. P. XIII. IN. AG. P. XVII. S. on a broken piece of white marble, O'So metre long and o-10 high, which is in all probability C.I.L. vi. 30033, which Hiilsen copied in the Vigna Amendola on the Via Appia; and the provenance of the fragments of marble, etc. still to be seen there is quite doubtful (cf. supra, 319). Returning to the road and going E. to the so-called Grotte Pallotta (Papers, i. 253, n. 3) I should add that Grossi-Gondi (p. 175) maintains that the inscription of C. Iulius Cornutus Tertullus (C.I.L. xiv. 2925) really belongs to this site, and thus gives us the name of the owner of the villa ; C C 386 THE BRITISH SCHOOL AT ROME. and certainly the testimony of Cod. Tusc. cit. I5ov (cited by Lanciani, Bull. Com. 1884, 210, and again in Storm degli Scavi, iii. 55) is not to be neglected. Further E. again is the Casale Ciuffa {Papers, i. 255) which stands upon the remains of an ancient reservoir : below it is another casale, also built upon a reservoir, but not marked in the Staff Map. For the district to the E. of this I must refer to my description already given in Papers, i. cit.

XIX.—FROM THE THIRTEENTH MILE OF THE VIA LATINA TO ROCCA DI PAPA AND MONTE CAVO.

We have already followed the road S. from La Pedica to the Ponte Squarciarelli and Marino {supra, 252), but we have not examined the district to the E. of it. Close to it, a little before it turns E. to descend to the bridge, I have noticed remains of a villa in opus reticulatum : the site is now occupied by modern houses. The Torre delle Streghe is spoken of by Tomassetti (142 n.) as though it was built upon the core of an ancient tomb, but no traces of ancient construction are now visible there (so also Grossi-Gondi, p. 193, n. 2). Here Stevenson saw in September 1891 a plain cippus of marble ; and at the Casino of Filomena Serafini, close by, a pedestal of white marble, of curious design, 070 metre in height, and widening from O'2O metre at the top to C23 at the bottom. The front was decorated with a head of Minerva in relief, with a winged helmet and Gorgoneion : the eyes were left hollow. On each side of it a bird's head projected. It was said to have been found in a vineyard further up the hill, apparently in the remains of a building of which the marble pavement was discovered, with the usual late burials under tiles. In the vineyard in which stood the so-called Torre delle Streghe the fragmentary tomb inscription C.I.L. xiv. 2552 was copied in the eighteenth century by Cardoni; but in the next century it was seen by De Rossi in the Vigna Santovetti on the way up to Rocca di Papa. Cardoni's idea that it contained the names of the consuls of 240 A.D. is incorrect. Not far from the Ponte degli Squarciarelli we may place the springs of the Aqua Julia1: a cippus of this aqueduct, bearing the number 302 {C.I.L. 1 Frontinus (i. 9) gives the distance from Rome as two miles to the right of the twelfth mile of the Via Latina, which takes us beyond even the Aqua Algidosia. But it is probably not to be CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 387

xiv. 4278; vi. 31563 b, Not. Scavi, 1887, 82) was found on the left bank of the stream, about 200 metres below the Cartiera, i.e. almost opposite the Abbey of Grottaferrata ; and another, with the number mutilated (C.I.L. vi. 31563 c\ Not. Scavi, 1893, 240) was found a little higher up, towards the second intake of the large mill ; but neither of them was, so far as is recorded, found in situ. There are now springs to be seen about 200 metres below the bridge in the bed of the stream, and Lanciani (Comentari di Frontino, 296) notes, as is still the case, that, while little or no water may be passing under the bridge, the springs themselves are plentiful (cf. Builder, xciv. 235, Feb. 29, 1908). Rocchi (Diss. Pont. Ace. Arch. Ser. II. vol. vii. (1900) 237 Jin.) speaks of masses of travertine which served for the intake, and states that he believes that he has seen the specus at three points. These springs may, I think, be safely attributed to the Aqua Julia. Holste {Cod. Dresd. F. 193, f. 4iv) considered the fountain of the Aqua Algido-ia (or, as it was called in his time, d'Angelosia) to be the real starting point of the aqueduct.1 There are, however, no actual remains of its channel to be

taken as an accurate measurement. The cippus no. 302 gives the length of the channel as 14J miles from the point at which it stood to Rome, while Frontinus gives the total length as 15J miles. 1 It may be well to quote the actual words of Holste in regard to the Aqua Julia {Cod. Dresd. F. 193, f. 41 sqq.): ' 1649, 14 Septembris fui Cryptae ferratae, unde ab homine locorum optime perito conductus fui, primum recta via ad orientem, qua curribus iter Marinum et Velitras ex Via Latina ad xii sive ad Cryptam ferratam deflectendo. D. C. circiter passibus progressis ad dictam viam occurrit fons a Card. Alexandra Farnesio Abbate dicti loci extructus, vulgo dicitur, la fontana dei Squarciarelli, et Crabrae ex proximo imminet cum hac inscriptione. Alexander Farnesius Card, aquam Tepulam ad publicum usum hoc extructo vase collegit M.D. Ixvii. Sed aqua eius fontis nequaquam eo, aut proximo loco scaturit, sed ductu subterraneo enra ad locum ac deinde ad dictum monasterium Cryptae ferratae perducitur. Sed nullum dubium est quin superioribus in locis concipiatur ex aliquo capite Iuliae. E regione huius fontis Farnesiani ad sinistrum latus Crabrae luculentae sunt scaturigines aquae limpidissimae, quas olim Iuliae partem fuisse dubium nullum est, imminet autem hisce scaturiginibus in colle vinea Thomae Vitacci, qui puteum profundum excavans in veterem aquaeductum incidit, per quem etiam nunc aqua cum strepitu labitur. A fonte Farnesiano paululum adscendentes defleximus tandem ex via ilia curuli, diverticulo sinistrorsum ad passus amplius mille, ubi ad sinistrum latus Crabrae insignis fons ex antro opera humana exciso profluit, slatim Crabrae iungitur. Vulgo la fontana d'Angelosia dicitur. Antrum illud in modum aediculae excisum fornice lateritio antiqui Romani operis clauditur, ostia quoque ipsa sive frontispicium antiqui operis lateritii vestigia ostendit. Hunc praecipuum Iuliae fontem caputque fuisse certum est. Paulo supra dictum fontem in planitie palustri concipiuntur aliae aquae scaturigines et in piscinam sive cisternam deducuntur, quae vulgo La Botte dicitur, et proxime sub hisce scaturiginibus nascitur Crabra loco valde uliginoso arundinetis consito. Primae illae scaturigines ex cisterna la Botte rivo subterraneo per cuniculos excisos in villas Tusculanas perducuntur, nee dubito quin hae scaturigines cum proximis illis Crabram olim constituerint. Omnes istae scaturigines oriuntur in territorio Molarae ad ipsum confinium territorii Abbatiae Cryptaeferratae. Crabra autem quae nunc quoque suo rivo Romam perducitur, duobus C C 2 388 THE BRITISH SCHOOL AT ROME. seen, and I do not know whether he was right in considering the brickwork at that fountain to be ancient, inasmuch as the fountain has now been enclosed.1 Secchi (p. 35 of the work cited supra, 228) records his discovery of an ancient channel, 10 to 12 metres above the level of the modern fountain of the Squarciarelli, which was running towards the Aqua Algidosia or del Canalicchio, and which he believed that he had found again lower down in the Vigna Passamonti, near the Villa Montioni {supra, 226) and near Fontana Piscaro (whether he is right in this state- ment I am not sure), and he notes rightly that this is the only water to which Frontinus' description would apply (i. 9): 'praeter caput luliae trans- fluit aqua quae vocatur Crabra. Hanc Agrippa omisit, seu quia impro- baverat, sive quia Tusculanis possessoribus relinquendam credebat; ea namque est quam omnes villae tractus eius per vicent in dies modulosque certos dispensatam accipiunt. Sed non eadem rnoderatione aquarii nostri partein eius semper in supplementing, luliae vindicaverunt nee ut Iuliain augerent, quam hauriebant largiendo compendii sui gratia. Exclusi ergo Crabram et totani iussu imperatoris reddidi Tusculanis, qui nunc forsitan non sine admi- ratione earn sumunt, ignari cui causae insolitam abundantiam debeant.' Holste is, however, perfectly right in pointing out that the springs of the Crabra take their rise in the Valle della Molara, as Lanciani, op. cit. 322, also says, at 612 metres above sea-level, under the Colle Bartolucci, to the N. of the 18th mile of the Via Latina: the water is now conveyed to Frascati by the Acquedotto Aldobrandino, constructed early in the seventeenth century {supra, 334) and restored by Canina in 1838-9 {Tuscolo, 85). Grossi-Gondi (p. 83, n. 2) unnecessarily supposes a misprint in Lanciani's text (cf. supra, 232 n. 2). At point 39s is a house (the Casale Iozzi, according to Stevenson) which occupies an ancient site : there is just below it a long low embanking wall of rough opus reticulatum of selce with quoins of stone, and square weep-holes at frequent intervals : beyond this again the front of the wall begins to be decorated with half columns. Near the house is a reservoir, and to the W. is another house, by which stands an ancient handmill. To constat ramis, quorum alterum ex valle Cryptoferratensi defluentem Maranam vocant, et rivum aquae Ferentinae Marino defluentem excipit paulo supra pontem ad decimum. Ultra Marianum nascitur Marinella.' 1 In 1901 the following brickstamps (from tombs 'a capanna' (?), but no further details were given) were found at the reservoir of the Aqua Algidosia, C.I.L. xv. 205, 211, 239, and a variety of 596 (Not. Scavi, 1901, 202). CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 389 the W. again, above the second ( p ' of' Giuseppe,' Stevenson saw the core of a tomb in concrete, 3"8o metres high and 270 square, parallel to the path which ran E.S.E. from the Ponte Squarciarelli to the Casalaccio,1 and which, from this point onwards, followed the line of an ancient road. At the Casalaccio there is an ancient reservoir with two chambers under the modern house ; and to the E. are clear indications of the site of a villa. To the S. is the Vigna Troili, in which De Rossi saw two fragments of marble, o'So metre high, with the remains of an inscription in letters o-22 high, belonging doubtless to a large tomb .... Cn.f. Pap{iria tribu) .... and .... eros [C.I.L. xiv. 2574). I saw the first fragment, but without the initial C, in 1909. I also noticed there a block of travertine 1 metre long and 059 metre high, with the upper half of three letters of a very large inscription (the upper half alone being C32 metre high) A D \ / / V i (The second letter was either a P or a B) and a marble cippus o-9O metre high, the inscribed surface measuring 0^54 by O'S6 metre, and the letters being 0^03 metre high, TI CLAUDIUS MNESTER CLAUDIAE PHILTATE ET CLAUDIAE VERECUNDAE CLAUDIAE SEVERAE F Tomassetti (p. 266 n.) notes in a vineyard on the left of the high-road an architrave of peperino with curious sculptures, and at the Casa Fondi, to the S.W. of the Villa Troili, the existence of various inscriptions, which had, however, been brought from Rome. To the S. of the Villa Troili on the modern high-road is the Casale dei Frati (formerly Villa Locatelli), which probably occupies the site of an ancient villa in opus reticulatum, of which much debris may be seen— though Stevenson {cod. cit. 5OV) was informed that the material came from Catorso {infra, 402) and that a male and female statue from the same site, which had been preserved there, had since been sold. At the Casale

1 Casasalaccio on the map is an obvious misprint. The lane between point 395 and the Casalaccio has been diverted in recent years. 39O THE BRITISH SCHOOL AT ROME.

Botti to the E. he noted architectural fragments, but saw no remains in the vineyard. The line of the path we have been following now falls into the modern road to Rocca di Papa: perhaps the path going on almost due S. and eventually reaching Palazzuolo is ancient, but there are no decided traces of this, for though it is sunk deep, the soil is soft, and I have only marked it as doubtfully ancient for a short distance. Before going further E. we must return to the district to the S. of the Ponte Squarciarelli. At the point of divergence of the modern road to Marino, I noted in May 1904 two heads placed in the wall, representing our Lord and the Virgin, the former mediaeval, the latter a Roman female head, with the hair parted in the middle and waved to each side with a veil over it. In the grounds of the Trappist Monastery of S. Giuseppe here two lead pipes were found in 1902, bearing the inscriptions . . . Antoni Albi and Deme- tr{ins) L. Ant(oni) AlbiJ) ser(t>us) feciii). 150 metres off were found remains of a villa in opus reticidatum and marble fragments (Grossi- Gondi 193= Bull. Coin. 1902, 320). To the S. of the monastery the modern road to Marino cuts through a drain and a reservoir in concrete: the latter is well preserved and has two chambers, each 3-90 metres wide and about 14*50 metres long, with three apertures each rio wide in the dividing wall. To the S. of this, in the Vigna Santovetti, are fragments of marble sculptures and column drums, which led me to suppose the existence of a villa here, but which may not really belong to the site. To the W., however, are extensive, though scanty traces of a villa, as also in the Villa Blasi (cf. Stevenson, Cod. cit. 51) and again further S.E., not far from the Casale Cavaletti. Rosa marks a villa at this last place. The path going E.S.E. back to the Casale Grotte dell' Acqua is deeply sunk, but the soil is so soft that it is not necessarily ancient. The Casale Grotte dell' Acqua occupies the site of a very large villa, to which belongs no doubt a wall in opus reticulatum of selce to the S.S.E. of the Casale, found in the recent excavations for the electric tramway to Rocca di Papa : on the N.E. side of it was a gutter of peperino slabs, covered in by a small concrete vault. Under the Casale itself is a large reservoir with two chambers, each 21*20 metres long. To the S.E. were many loose pavingstones, which may have belonged CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 391 to a road leading to another villa which was brought to light in the course of the tramway works. It is in rough opus reticulatum, with some brick- work : the floors of two rooms were visible, one in plain cement, the other in cement vith fragments of marble placed in it at intervals. Under one of the floors is a vaulted reservoir, and most of the building is on vaulted substructions of irregular orientation and shape. To the E. of it was found a lead water pipe, without inscription {Not. Scavi, 1905, 272). I saw one brickstamp {C.I.L. xv. 1121. a—first century, A.D.). To the S.E. of this villa in the tramway cutting are the scanty traces of another building; and a little beyond this, in the Valle Scura, is the, lower station of the funicular railway on to R'occa di Papa itself. The high-road offers no definite traces of antiquity, and the ancient road may have passed further S.W. : I have marked it conjecturally as following the older path which preceded the zigzags of the modern road. To the N.E. of the Grotte dell' Acqua, on the further side of the high- road, is a vineyard in which are the remains of another villa : there is a large platform in opus reticulatum facing W., and on the top of it are mosaic pavements.1 To the N. is the Casale della Morte, and to the N. again the Vigna Portone Gazzi, where there are mediaeval ruins: neither of these names appears on our map. To the N. of the last is the Vigna Pio Vitale, at the Casale of which are two well-preserved column bases of tufa and some blocks of tufa. For the district further N. see infra, 402. I have doubtfully marked as of ancient origin the path ascending S.E. from the neighbourhood of these houses to the high-road : the evidence is by no means sufficient for certainty on the point. To the E. of this path is an aqueduct crossing the Vallone Arcioni in a S.W. direction, which takes its name from these remains, of which I have not found any full description in any other writer. Nibby {Analisi, i. 118) attributes them to the Aqua Algentiana (which is merely a false reading for Alsietina), while Canina, Tuscolo, 85, wishes to connect them with the springs from the Valle della Molara {supra, 388). I give a photograph of the central portion in Plate XXXIII. Fig. 2, and a detail of the construction in Plate XXXV. Fig. 4.

1 The site of the discoveries described in Not. Scavi, 1878, 260, as having been made in a vine- yard belonging to Locatelli, among the ruins of an ancient building, some 3 kilometres from Rocca di Papa, on the left of the high-road, is not exactly clear ; but I suspect it may be this villa. The ruins are described as consisting of a long line of subterranean corridors lined with white cement: within them were found some statues, which are there described. 392 THE BRITISH SCHOOL AT ROME.

There are seven arches standing, three incomplete on the N.E. side, and two or three on the W. The highest, over the stream, is about 8 metres high with a span of 4 metres ; some of those on the west bank have a span of 4/20 to 4^50 metres. The construction is of peperino concrete faced alternately with baked bricks 0x53 metre thick and small rectangular blocks of peperino C04 to O'o6 metre thick ; the mortar courses average O'O2 metre thick. At an interval of 140 metre at the bottom of the piers, but as little as C53 metre higher up, a bonding course of tiles C04 metre thick runs right through. The arches are of tiles, about one voussoir in every.four going right through the width. The piers are 2-io metres thick and 2-oo wide, and some of them have auxiliary buttresses on the inner side. I could see no traces of the specus, nor could I follow the aqueduct E. or W.: there are no traces of deposit to be seen, and its absence renders it impossible to determine the course without excavation. From the level it might well be going towards the reservoir of the Grctte dell' Acqua. The valley goes on S.E., becoming deeper, until it reaches the great cliff below the Pentima Stalla, from which it takes its origin. On the further side of it, S.W. of the letters ' ciole ' of the word ' Spadacciole,' Stevenson has marked the pavement of an ancient road, which still exists, running S.E. up the hill through the wood as far as the Via delle Molare. Just before it reaches this, on its N. edge (about the first' e ' of ' delle '), are the ruins of a villa, which Rosa, it seems, also saw. They seem to have been excavated not very long ago, and I am inclined to identify them with the villa in which were made the discoveries described by De Rossi in Ann. Inst. 1873, 178 : this was situated within the macchia of Rocca di Papa, and not far from the aqueduct (infra, 394) of which he speaks in the preceding pages. In this villa there were found the lower part of a bas- relief of the third century A.D. representing three female figures with a dedication to Fortune S{anctae) D(eae) Fiortunae) F(elicitati) d{onuni) d(edit), the name of the dedicator having perished with the upper part of the relief (C.I.L.xiv. 2568), a sepulchral inscription of the second or third century A.D. D(is) M{anibus) Sex(to) Antis(ti)o Montano qiiii) v(ixif) ainnis) duobus m{ensibns) II diiebus) F( ) Festianus actor b(ene) m(erenti) (ib. 2569), which has, but without sufficient reason, been taken as an indication of the name of the owners of the villa, four lead pipes, of which two, C.I.L. xv. 7834, xii. Cn. Cassi Euscatij), 7848 M. Trebelli Hieroclis, give us, probably, CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 393 the names of persons to whom the house once belonged, and two others, ibid. 7858, T. Cispius Verns fec{if), 7867, Septimius Secundinus fecit, merely give us the names of the makers of the pipes; with the pipes was found (and destroyed) a leaden box which served for the distribution of the water and from which these pipes would have begun their course. Among the bricks found in the building were some bearing the stamps C.I.L. xv. 533. 8 (M. Aurelius or Commodus) and 2370 (first century A.D.?). Lanciani {Not. Scavi, 1878, 259) describes statues of Diana and of Juno (?) preserved at Rocca di Papa, and said to have been found here in or about 1870. A fine statue from the Selva di Rocca di Papa, in the Locatelli property, representing an ephebus or a youth exercising in the palaestra, was, it is reported, removed to Frascati in 1884 : it was half life-size, with the arms and one foot missing {Not. Scavi, 1884, 348). Whether it was found in these ruins, or in those mentioned supra, 391 is not clear from the description. I cannot trace the pavement of this road further in either direction : there are traces of antiquity in the Via delle Molare—just W. of the ' M ' is a cutting with pavingstones in it—tout only at that point: so that whether we should suppose that an ancient road descended to the Madonna della Molare or to the Casa dei Guardiani is uncertain. To the S.E. lies the village of Rocca di Papa, which has been thought by topographers to occupy the site of Cabe or Cabum, the city of the Cabenses in Monte Albano, named by Pliny {N.H. iii. 64) among the peoples of Latium who had perished, who, as Mommsen remarked {Bull. Tnst. .1861, 206), are identical with the Kaftavoi of the list of the thirty cities of the Latin League (Dionysius v. 61): their name, as he conjectured, probably survives in that of the modern Monte Cavo,1 and was still preserved in imperial times as that of a priesthood {Cabenses Sacerdotes feriarmn Latinarum motitis Albani, C.I.L. vi. 2i73 = xiv. 2228, ox Sacerdotes Cabenses Montis Albani, ib. vi. 2174, 2175). Whether the name Papa is not also a corruption of the same is a more difficult question, though it seems possible. Tomassetti, however (p. 266), does not accept the idea ; and it does not appear before the end of the twelfth century. Lanciani {Not. Scavi, 1878, 259) reports the discovery of some remains in opus quadratum at the entrance of the village, at the corner of the road leading to Ariccia.2 I cannot agree 1 Cavi or Cave is the correct form of the name—in a document of 1249 we have a mention of the via silicata Montis Cavae (Tomassetti, 280). 2 The brickstamp C.I.L. xv. 1068. a. 17 is recorded as found at Rocca di Papa in 1866 : cf. 1324. 3 (near Rocca di Papa). 394 THE BRITISH SCHOOL AT ROME. with Tomassetti in believing that any of the wall under the foundations of the now destroyed mediaeval castle above the village is antique :1 the blocks are too small and are clearly mortared. Nor is he right in placing this castle on the hill marked 731 metres in the map: it really occupies the height marked 760, while the discovery made by M. S. De Rossi must be placed, not near it but some way further N.E., probably, from the map (Ann. Inst, 1873, p. 169 and tav. d'agg. R.S.) on the Colle Molare. Here, in a cutting of the forest, near the edge of the crater, there was seen and partly demolished a portion of an ancient enceinte in opus quadratuni, in which the break for a gate or door was seen. Close by an archaic tomb was found, with archaic pottery in it, with a rough stele of local stone; and neolithic weapons were common. Pinza too (Man. Lincei xv. (1905) 35) mentions the discovery of a bronze celt in the Campo d'Annibale. I have not myself seen anything of the wall; nor have I been able to hear of the existence of any of the five cippi of the Aqua Augusta (De Rossi, Ann. cit. 170 sqq. ; C.I.L. xiv. 2567), which, from the map already cited, must have been found a little further N.E. ' in cutting the underwood which descends from the Campo d'Annibale (the name popularly but groundlessly,2 given to the inner crater of the Alban volcano) towards the valley of Algidus and the Via Latina' and ' in the part of it which descends from the springs of Pentima Stalla to the fields of Rocca Priora.' The aqueduct itself was not found, and the cippi throw no light on its origin or course, though De Rossi (Ann. cit. 177) wishes to connect it with that in the Vallone Arcioni, of which we have spoken supra, 391. See Lanciani, Comentari cit. 327. To the W. of the Colle Molare a path goes off N. : it first runs down gradually, through a cutting 470 metres wide and some 40 deep: this is mentioned by Nibby in his article on Le Vie degli Antichi, 42 (in Nardini, Roma Antica, ed. iv. (1820) vol. iv.). It then descends steeply through the Bosco Riguardata Molare, but through a shallower cutting. Excavations made in 1793 3 by Prince Filippo Colonna in the Bosco Riguardata in a circular building of about 150 feet in diameter with eight niches, supposed to be a ruined temple, led to the discovery of the statue of

1 He gives a description and plan of it pp. 273 sqq. 2 That the name had no classical foundation was seen by Fabretti (De Aquis, Rome, 1680, p. 181, §§363 sqq.); and Tomassetti (p. 280) points ovit that it is taken from that of the Annibaldeschi, the mediaeval lords of Rocca di Papa. 3 The precise date is given by Fea in his MS. notes in the library of the French School at Rome (I. 12, f. i85v). CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 395

Artemis now in Berlin (Beschreibung, no. 59) which was restored by Pacetti {Memorie Enciclopediche, iii. 90, no. 55)- Niebuhr's statement that it was found near Le Frattocchie is unsup- ported by other evidence. See Levezow, Berliner Kunstbl. 1828, 317, who gives no provenance. I have never actually seen the building and cannot certainly identify it. The only ruins I have seen thereabouts, and these do not correspond to the description, would fall about where is the second ' a' of ' Riguardata,' and belong to a building some 20 metres long by 10 wide, with fallen vaulting in tufa concrete : the mortar is bad, but I could not see any certain indication of date anywhere. At the Casale dei Guardiani the path falls into one leading S.E. with which we shall deal later {infra, 403). Just at the point where the path we have described leaves the Via di Rocca Priora, in the angle between the two paths, there is ancient debris— brick, pottery, etc. The Via di Rocca Priora, when it leaves the crater, descends sharply : in places it is paved with large rough blocks of selce, probably of ancient origin : where they are gone the path is worn away deeper. Some way down it begins to be cut through the peperino, and then measures 2-58 metres from edge to edge, and r8o metre between the ruts : the cutting is not more than five feet deep, and there is a drainage channel cut down the middle. It crosses the path going S.E. from the Casale dei Guardiani, and falls into the Via Latina near the Muracci {infra, 412). A third path from Rocca di Papa, which falls into the path from the Casa dei Guardiani, is the Via delle Selve, which crosses the floor of the inner crater of the Alban volcano, passes N. of the Monte Vescovcf, and crosses the rim of the inner crater at the Forcella. Of its antiquity there is no certainty, but it would have formed an important—almost a necessary line—of communication.1 There is, on the other hand, evidence for the existence of an ancient road ascending to the summit of Monte Cavo from this side. Its precise course as far as point 685 (on the E. edge of Map II. of Papers, iv.) cannot be fixed, but it probably followed the line of the track running S. from point 760, and I have unfortunately omitted to mark it conjecturally. From point 685 it must have ascended rapidly, butwe only reach actual

1 Under the N. slopes of Monte Cavo, on the S. (or rather W. )edge of the Campo d'Annibale, three small Egyptian statuettes, two of porcelain, one of blue stone, were found in 1885. {Bull Inst. 1885, 182.) 396 THE BRITISH SCHOOL AT ROME. traces of pavement further up, to the E. of point 790, almost due W. of the summit of the mountain : here it joins the road ascending from the S., and both proceed towards the top together. The modern path up the mountain from Rocca di Papa avails itself of the last portion—that which is included in Map I. of the present volume— which begins shortly after this junction : the road runs first due S. and then reaches the top by a wide bend. The pavement is well preserved on this ascent (see Plate XXXIV. Fig. I) and is 2-6o metres (9 Roman feet) in width. On some of the pavingstones in this last portion of the ascent the letters N and V are found often on adjacent blocks (N occurs seven times, V five or six times): Mommsen and Hlilsen explain them as meaning N(ova) and V(etus), i.e. as marking parts of the road that had or had not been repaired {Rom. Mitt. 1890, 70 fin.). W\bby, Analisi, i. 114) had already noticed them, but had interpreted the letters to mean Niuminis) V(id). He remarks that one can still see the roughening on the surface of some of the pavingstones, to prevent slipping. We must return for a moment to the road ascending from the S., though the greater part of it falls outside the limits of the territory we have undertaken to describe. There has been a good deal of discussion as to the course taken by the ancient road from Rome to the summit of the Alban Mount—it is often called the Via Triumphalis by modern writers, though there is no authority for the name. Nibby at one time thought that it diverged from the Via Appia near Bovillae and passed by Marino, and then turned due S. near the Ponte degli Squarciarelli ( Viaggio Antiquario, ii. 74) : later he made it follow the N. edge of the lake of Albano, and so reach Palazzuolo (Ana/isi, iii. 114). Rocchi (Bis. Pont. Ace. Arch. Ser. II. vol. vii. (1900) 223 sqq.) prefers to make it diverge from the 12th mile of the Via Latina {supra, 227, 280). He mentions the existence of pavingstones in situ and in the fieldwalls just at the pilasters of the entrance gate of the old Villa Colonna (1629-1640), i.e. at the point where the main track diverges E.S.E. from that which goes to Palazzuolo, and of two crepidines in tufa some 625 metres from the first traces, which gave the width of the road as 8 Roman feet. To the E. of Palazzuolo more of its pavement is preserved in situ ; and here to the N. of the road is a large reservoir of selce concrete, measuring some 26 by 21 metres and open to the air. The road ran E.S.E. as far as the Casa dei Guardiani, which is just outside Map CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 397

I. of the present volume on the W., and is built over an ancient building in brickwork of the second (?) century A.D. (the bricks being O-2I metre long and O'O35 thick, and the mortar o-oi5 thick), of which two vaulted chambers remain. Two column drums of gray marble, O'42 in diameter, are to be seen there; and in 1897 a statue representing a Roman matron, of the well-known ' Pudicitia' type, was found in digging near the S.E. side of the Casa (Not. Scavi, 1897, 389). Close by (if the two vaulted chambers described are the same, and if, as is probable, the Casa was erected in the interval) remains of a bath estab- lishment were found in 1894, consisting of two almost semicircular chambers, one paved with mosaic, the other with white marble and porta santa : the rooms were heated by hypocausts, and on the large tiles, O'6o metre square, which rested on the pillars and supported the floor were the stamps C.I.L. xv. 171 (about 138 A.D.), 207, 674 (both 134 A.D.): cf. Not. Scavi, 1894, 405. It is a pity the later report does not mention the earlier. Thence the mediaeval post-road (which probably followed as before an ancient line) continued past the Fontana Tempesta to the pass of the Faiola and Velletri, leaving Nemi on the S.W. ; while the road up the mountain (which was joined by a branch from the S.W. at the Casa dei Guardiani (but passing N.W. of it, not S.E. as the modern path does) form- ing a mode of access both from Ariccia and from Genzano, in all probability1) turns off first to the N.E., and then, as it ascends, runs N.W.2 and then due N., until it joins the road coming up from Rocca di Papa, E. of point 790 (supra, 396) as may be seen in the map : its paving is still traceable for nearly the whole distance, and is in places quite well preserved: Chaupy (op. cit. ii. 115) saw it almost entire : the first portion was destroyed in his time. It is recorded that Pius II., when he saw a workman damaging it, gave orders for its preservation, and that Alexander VII. drove up to the top of the mountain by it in a carriage (Mattel, Tuscolo, 18). The Alban Mount (Albanus Mons, To 'AXftavbv opos) is, strictly speaking, Monte Cavo itself, but the name has been extended to the whole massif of both craters of the Alban volcano (Papers, iv. 9). It is the highest point of the Alban Hills (956 marked on the map as the altitude of the

1 The traces of this branch are hard to follow, but can be found in the wood as far as the modern road from Ariccia to Rocca di Papa. It is 2p8o metres wide : Stevenson's map first called my attention to it. On its E. side are remains in opus rcticulatum. 2 Where it turns N. yet another branch may have joined it; but this must be discussed else- where. - 398 THE BRITISH SCHOOL AT ROME.

Maschio delle Faete to the E. is probably an error for 936).1 It was in ancient times crowned by the temple of Jupiter Latiaris, the federal sanct- uary of the Latin League, at which in the spring of each year the festival of this ancient confederation was held {feriae Latinae). Here too those generals to whom the honour of holding it in Rome was not conceded celebrated their triumph. In imperial times the worship was carried on by the Sacerdotes Cab- enses montis Albani {supra, 393). As a sketch in Cod. Barb. Lat. xxix. 215 f. 38 shows (it was published by De Rossi, Ann. Inst. 1873, p. 167, tav. RS), even in the seventeenth century there remained only the foundations of the temple, of massive blocks of peperino (I measured one 2"65 metres long by 070 thick, another was 1-90 by C83 metre) and unimportant architectural fragments, which were used in the erection of the Passionist Monastery by Henry, Cardinal York, and may still be seen in the walls of the garden : that he destroyed the well-preserved remains of the temple for his purpose is incorrect, and so is the date 1783 or 1788, given by some writers. Hiilsen (in Pauly- Wissowa, R.E. i. 1310) gives the date as about 1777, but I think it should be a little earlier ; for Chaupy {Maison de Campagne d'Horace (1767) ii. 114) says ' les vestiges du Temple de Jupiter-Latial ont ete fort sensibles jusqu'a nos terns, mais diminues par un hermitage que se fabriqua sur Monte Cavo un Francois de la Cour de cette Reine de Pologne qui se retira a Rome, ils sont sur le point de disparoitre par la maison et l'eglise, que la nouvelle Religion des Passionanti s'y batit.' He then alludes to C.I.L. xiv. 2242-4 : ' il a ete tire d'un endroit ou Ton apenjoit les fondemens d'un grand Portique qui entouroit le Temple.' Seven years earlier Piranesi {Antichita di Albano, Rome, 1762, tav. i., ii.) gave elevations and sections of a Doric entablature and column which no doubt belonged to the temple, and also of a larger cornice : Canina, Edifizi, VI. tav. lxxii. attributes the larger cornice to the temple and the smaller fragments to a colonnade surrounding it, and gives a quite arbitrary reconstruction of the plan. Delbriick {Die drei Ternpel in Forum Olitorium, Tav. iv. 2) reproduces Piranesi's engraving of the entablature and column. In the Attidel Camerlengato, Tit. iv. fasc. 43 are some papers in regard

1 The same applies to Monte Peschio on the rim of the outer crater, which is really 939 metres above sea-level. See Papers, iv. 9, notes I and 2. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 399 to the preservation of the remains in the first half of the last century. The correspondence begins with a request, dated 30th July 1824, from the Provincial of the Passionists, for leave to make a superficial excavation for blocks of stone within their enclosure for the campanile of the church, which had to be transferred elsewhere. The request was referred to Carlo Fea, who strongly objected, and leave was therefore refused. The request was renewed in February 1825, the Provincial pointing out that he only desired to dig in the garden and use the stones that were found, and not to touch those which formed part of the temple: the Commission was again inclined to refuse permission, but decided to visit the spot first. The result of the visit was a definite refusal, the Commission deciding that there was plenty of other material available. A sketch and plans of the new campanile by Valadier are given. In August 1836 the Cardinal Chamberlain made further inquiries to see if his orders had been respected, but the results are not here to be found. Excavations were conducted on the summit of the mountain by the German Archaeological Institute in 1876, under the direction of Michele Stefano de Rossi, and are described by him in Ann. Inst. 1876, and tav. d'agg. Q. They rendered it possible, first of all, to assign the drawing just mentioned to its proper place on the site. It showed, if it is to be considered as entirely trustworthy (though it was not found possible to ascertain the correctness of the details), that the temple was a rectangular chamber, with an apse on the N. side, placed near the N.W. angle of the sacred area, which lay to the E. of the modern monastery buildings, and to which the road ascending the mountain led directly. To the W. of the temple was a large construction, supposed to be the base of an altar. In this area was a shaft leading into a passage which served for the collection of rainwater, leading to a large cistern from which the water was distributed by another passage, not cemented, but floored with selce, in which, therefore, lead pipes must have been placed. Fragments of plates of lead with inscriptions (C.I.L. xiv. 2233-2235 = xv. 7808-7810) were found in the cistern, which were connected with the distribution of the water, belonging in all probability to the receptacles from which the pipes originated. One bore the inscription c\ur{ator) aed{iuni) s[acrarum ]ms. Ca ... .in another the numbers cxli., cccciiii., with the consular date 11 B.C., and a third the numbers cccxlv., cxv., 400 THE BRITISH SCHOOL AT ROME. the significance of which is not clear : the numbers are certainly too high for guinarii to be meant. De Rossi rightly calls attention to the fact that in n B.C. a law was passed about the distribution of water (Frontinus § 99 sqq?). To the S. of the area was a room with pavement of black and white mosaic, with a drain for rainwater to the S. of it. I should imagine that it was here or in the irregular walls (no. 11 of De Rossi's plan) that the various brickstamps were found which, though not mentioned in his report, are recorded in C.I.L. xv.1 Among the objects attributable to the stips votiva of the temple were some specimens of archaic pottery and of aes rude. The most important of the portions known to us of the fasti of the temple relating to the feriae Latinae had been discovered previously, some in 1765, some in or about 1869, and only a few fragments came to light: they were found in the cistern. All were cut on blocks of marble, which formed, it seems, the walls of the temple itself. They began to be inscribed in the time of Augustus, but seem to have commenced from the time of the Decemviri {C.I.L. xiv. 2236-2244). Other inscriptions found at the top of the mountain are—a dedication to Jupiter Latiaris, made by L. Rubellius Geminus, consul in 29 A.D., carved on a fine marble base (ib. 2227), a dedication to Antoninus Pius (?) made by the Sacerdotes Cabenses (ib. 2228, supra, 393), a fragment of a cippus with the letters Imp. (ib. 2229), another fragment of a dedication to Agrippa as patronus in his third consulship (13-12 B.C.) (ib.. 2230), a base of peperino bearing an inscription which Henzen read \resi\itue\runi\ Ardeates (ib. 2231), but which Hiilsen has made out from a re-examination of the original to be Divei Ardeates: he attributes it to the fifth century of Rome, and considers it thus to be one of the oldest Latin inscriptions extant (Rom. Mitt. 1895, 65). Whether the fragment ib. 2232 aedem vei\ustate c\ollapsam resti\tuit

1 193, 6, 194, 2 (early years of Septimius Severus) 204, 23 (Severus) 226, 12 (M. Aurelius) 276 (115-120 A.D.) 367, 7 (M. Aurelius) 399, 14 (Faustina the younger) 400, 10 (do.) 435, 4 (Severus). 550, a. 15 (about 123 A.D.) 563, i. 37, p. 54, A. 79 (123 A.D.) 578, b. 10 (Hadrian) 585, b. 28, d. 59 (Hadrian) 596, 24 (Hadrian) 602, 4 (Severus) 737, 4 (161-168 A.D.) 754, a. 30 (Antoninus Pius?) 767, 6 (Severus) 774, 2 (Severus) 883, a. 3 (middle first century A.D.) 904, f. 27 (Trajan) 923, 3 (beginning of second century A.D.) 1012, a. 1 (after 108 A.D.) 1068, a. 18 (145-155 A.D.) 1086, 16 (154 A.D.) 1116, c. 8 (123 A.D.) 1135, 3 (end of second century A.D.) 1210, 15 (127 A.D.) 1609, 10 (after Diocletian) 1838, a. 3, d. 8 (Hadrian) 2249. To these we may add ib. 221, a. 8 (Faustina or Commodus) 400, 9 (Faustina the younger) 1058, 8 (125 A.D.) found in the eighteenth century ' in the ruins of the temple of Jupiter,' a statement to which we need not give credence, while 403, 7 (Commodus) was found on the mountain according to Riccy {Alba Longa, 70) and 873, a. I (first century A.D.) and 1146, a. 9 (Antoninus Pius) were found about 1871. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 401

which was seen by Armellini {Cronichetta, 1876, 139) in Rocca di Papa, can refer to the temple itself, as he thought, or whether indeed it was found on the mountain, is uncertain : it may refer to any building that needed repair. The fragments ib. 2245-2250, 4210, a, b, were found on the mountain, but cannot be assigned to the fasti, and to what inscriptions they belonged is uncertain : 2249 seems to be a portion of a ritual poem, and possibly ib. 4210 c may belong to it. The exact provenance of Kaibel, I.G. xiv. n 18, recorded by Fabretti as ' below the Alban mount on a rough altar' is unknown, and it may or may not have belonged to the temple. Its text is Ail icepavvlw. Among the inscriptions relating to the Alban Mount, though not found here, we may note C.I.L. vi. 14844 discovered in 1736 on the Esquiline near the columbarium of the Arruntii {ibid. 5931 sqq.). Clange Hilarionis v(ixif) a{nnis) xx perit in Monte Albano xiiii Kal. Iul. L. Arruntio M. Lepido Cos{s). {ijth June, A.D. 6).

XX.—THE VIA LATINA FROM THE THIRTEENTH TO THE TWENTIETH MILE. Beyond the thirteenth mile the ancient road followed the modern line for a little way, until it was joined by the branch which comes E.S.E. from the Frascati road {supra, 257): then it must have cut across the bend E. of the Casale Reali, where there are many pavingstones in the fieldwalls : then it passed S. of the modern road. At the point where it crossed the Fosso dei Ladroni, and was running io° N. of E., scanty remains of a substruction wall in polygonal work of porous selcex (' madre di selce'), with opus quadratum on top, are preserved, and there are also supporting walls in concrete ; but I do not think there can have been a bridge there in ancient times.2 A branch road seems to have diverged from it S.E. A little after the crossing a path turns off S. towards Rocca di Papa : it soon bifurcates, but neither branch seems to be ancient. Close to the house marked 415 I saw brick debris, and there is a reservoir to the W.S.W. towards the Acqua Algidosia {supra, 388), the so- 1 The blocks are well jointed, and are about 040 to 060 metre high and 0-85 metre deep, but the faces are rusticated. 2 This portion of the ancient road was known to Eschinardi : in Esposizione, cit. 410 (Venuti, 282) he notes the discovery of two large jars, which he believed to be ossuaries, at the side of the road. D D 402 THE BRITISH SCHOOL AT ROME.

called Grotta delle Streghe, nearly 40 metres long and 3 wide : to the W. of it are walls in opus reticulatum, belonging to a villa in which late burials were found, and also a fragment of an inferior draped statue. To the E., just S. of the ' a' of ' Pratone,' Stevenson indicated the ruins of a villa in the former Vigna Luccichenti: there is a small reservoir in opus reticulatum of selce at the ground level, and to the W. of it much debris with fine marble, while to the S. is a long substruction. To the S. in the path going E. there is a conduit, cut through by it, formed by a line of terracotta pipes croa. metre in diameter inside, enclosed in a channel of small blocks of stone 0*25 metre wide; and to the E. of this, just where the communal boundary line turns due S. for a little while, W. of the legend ' V. Catorso,' is a reservoir about 20 metres long, with two chambers, on the E. edge of a large cornfield. To the S. again, under the house marked 447 in the map, is a large vaulted chamber which, however, seems to be mediaeval. To the E. of this house is the Vigna Giovanetti, at which is still preserved the inscription C.I.L. xiv. 2494 Philerois) \_et] Antiochu\s\ Laribus d[onuni) d{ederunt) liibentes) mieritis). It is upon a cippus of tufa, C45 metre square and o-55 high as far as the base moulding, which is o-22 high, the base being o~S7 square. The letters are deeply cut and good, belonging to the first century A.D. Those of the first line are o"ii metre high, and the rest respectively o-o8s, O'OJ, and C075 metre high. I may note that the o at the end of the first line is very small owing to a miscal- culation of the space available, and that some of the other letters are cramped. To the E. again, E. of the letters ' Sp' of ' Spadacciole,' is a small house, to the W. of which is the walling of a cistern of opus reticulatum with tufa quoins and three courses of brickwork below. There are numerous pavingstones, which might belong to the road which we saw further S. {supra, 392); and though I could obtain no information as to its course, this seems to agree fairly well with what Stevenson was told {Cod. cit. 6y). In that case it would have been taken to mark the boundary line which still exists, between the cultivated area and the woods. Further N.N.W. in the wood (W. of the ' B' of' Borghese ') I noticed other debris, and Rosa also saw ancient buildings here. The Via Latina, to which we must now return, passed immediately to the S. of the Casino Le Molare: then it ran straight on due E. and must have recrossed the ravine which follows the S. side of the modern road, and CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 403 have taken more or less the line of the latter, though there are no certain traces until we reach the Osteria di Molare, just beyond which there is some pavement in situ 4/15 metres wide, while further on, between point 553 and the sixteenth mile, pavement may be seen in the stream bed. A little after the fourteenth mile are the remains of a villa on its S. edge, in which there is a good example of facing in rough irregular blocks of selce, illustrating the transition from the older polygonal style to opus incertum from which opus reticulatum was developed. Cf. Melanges de I'licole Francaise, xxv. (1905) 183. Above it is a reservoir with two • chambers. Hence a cutting runs E. to the N. of the hill marked 516 in the map, which does not seem to be that of an ancient road. Further S., on the W. slope of the hill which is known as the Colle della Tartaruga or Colle Precolio, and on the E. edge of Map II. of Papers, iv. are'the remains of another villa: the platform faces W.S.W., and four shallow arcades of opus incertum (almost reticulatum) of selce are preserved : the arches have voussoirs alternately of selce and peperino, and in front of the N. end is a later addition ; but the whole was apparently quite a small building. A specimen of the construction is give in PI. XXXV. Fig. 3. We now pass to Map I. of the present volume. On the top of the hill (point 516) Stevenson marks the existence of something ancient, but there is only a little debris ; and the same is the case below the chapel of the Madonna delle Molare,1 where there is brick and other rubbish, including a half column of tufa. Past this runs a track, leaving the high-road at the Fontanile S. Nicola, and running S.E. to the Casa dei Guardiani and on beyond it. For a long way there are no certain traces of antiquity,2 though in places it is deep cut, and it is significant that it follows the boundary line between the forest and the cultivated land and between the communes of Rocca di Papa and Rocca Priora ; nor are there any remains on the rim of the inner crater, which rises on its S.W. side.3 Just before it reaches a

1 The Castel di Molare is entirely mediaeval (Tomassetti, 282 sqq.—a photograph is given in his Campagna Romana, i. (1910), p. 178, Fig. 70), but there is some ancient debris : a hoard of mediaeval coins was found near it in 1902 (Bull. Com. 1902, 327). 2 The road from the Madonna to the Casa is sunk deep, and about five or more metres wide : but the soil is soft, and the pavement now visible is mediaeval. The track S.E. from the Casa to Colle Tondo, which I have marked doubtfully, is also deep sunk, and there are many bits of loose selce. 3 Somewhere near this road (between Rocca di Papa and Rocca Priora) was found the brick- stamp, C.I.L. xv. 1211. 3 (about 123 A.D.). D D 2 404 THE BRITISH SCHOOL AT ROME. branch coming from La Forcella {supra, 395), the actual crepido is preserved in situ, running S.E., and many pavingstones may be seen lying loose along its line, and also a few in situ. Here are the remains of a villa —I saw a chamber in opus mixtum, with a modern hut (' capanna') built over it, in which was a brick drain C48 metre wide, and was told that other rooms had been found, but covered up again. The land belonged, I was informed, to Giovanni Giovenazzi of Rocca di Papa. Beyond this point the path shows no further traces of antiquity : a little before it reaches the Riserva di Domatore it is joined by the main path from La Forcella, and then by another path from Rocca Priora {infra, 412), and shortly afterwards falls into the path along the Valle Vivaro, which leaves the Via Latina to the W. of the Lago della Doganella {infra, 413). This has no natural trace of antiquity, but is probably an ancient line of communication. At the point some way to the E. marked in the map, E. of the Fontanile del Ributto, W. of point 681 (the Colle del Favo) I have seen brick debris, and Stevenson {Vat. Lat. 10572, 34V) notes ruins of opus reticulatum near this hill. He tells us that he was informed by Nicola Santovetti of Grottaferrata that a torso of a statue, then at the Fontana del Vivaro was found near the ancient road {i.e. the Via Latina) ; and that there was a spring at the Fosso della Vetica, which was very scanty in summer, but copious in winter. At point 585 (II Meleto) on a hillock many lead pipes with inscriptions had, as Stevenson was told, been found fifty years before,, and he had himself seen much debris there : I saw brick rubbish at the point marked in the map a little to the E. A tomb had been, he says, found in 1891, 'a capanna,' i.e. with tiles forming the roof (unless indeed a hut urn is meant) with a mirror, and not far off was found a helmetx; and further S.W. just beyond the limits of our map, another tomb was discovered. We must now return to the territory on the N. side of the Via Latina, which we have so far omitted. Not far from the road, due S. of the theatre of Tusculum, and W. of a fieldwall and fence, are a few blocks of a polygonal wall in selce approximating to opus quadratutn, facing S. ; against it was built a drain in concrete, with a pointed roof: some of the

1 The exact words are ' Sep(olcro) a capanna con specchio presso Monsignore( ?) trov(ato) nel 1891, e vicino fu l'elmo (?). A Valle di Pratone al Vivaro dove ci e la parola ' di' di ' V. di Vivaro' [the lettering has now been altered slightly] ivi pure altro sep(olcro).' CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 405 walling under this drain is of opus reticulatutn. To the N.W. is a large platform, running about 2OC W. of N., and to the N. E. of the platform is a fallen fragment of the concrete wall of a reservoir, and also a building, possibly a tomb, of which I give a plan: the inner chamber is all of good brickwork, but the outer is entirely of opus mixtum and has probably been added later. The stairs led to an upper story, which is not preserved : the building is built into the hillside and '""" •• > t r 1—. only stands free in front. To the E. of the field wall are traces of polygonal walling (of a platform) and above there is a reservoir consisting of a single vaulted chamber in concrete, 15 metres long and 3*50 metres wide. Further to the N.E. on the path leading up to Tusculum under the slopes below the citadel are remains in concrete, indicated in the map. Further E., just after the fifteenth mile, a path diverges to the N., which follows the line of an ancient road, ascending sharply. To the E. of it are the remains of a platform of rough polygonal blocks of selce, preserved only to a height of one or two courses : the blocks are about O"9O metre deep. Just above, on the edge of the road, and cut through by the modern path, are the remains of the platform of a villa in opus reticulatum of tufa and selce, with stone quoins : there are four shallow arcades in front, and behind are extensive substructions. I should imagine this is the site referred to by Canina {Tuscolo, 108) as excavated in 1834: remains of baths and hypocausts were found. To the W. of the head of the valley too are the ruins of other buildings, including a large reservoir some 20 metres long ; and to the N.W., due E. of and below the highest point of the arx of Tusculum, are remains in opus reticulatum. The path we have been following then passes through a cutting1 in the gap between the arx and the hills to the E. of it, and arrives at a meeting-point of paths—one, probably of ancient origin, though certain traces are not visible, going W. to the road to Camaldoli {supra, 371), another descending N. to Monte Porzio {supra, 377), for which there is no definite evidence, and a third going E. along the ridge, which is certainly ancient, and which we must now follow. The paving was well preserved within living memory, and there are

1 The cutting is a fairly deep one, and it is conceivable that in earlier times it was a ditch connected with the defences of Tusculum. 406 THE BRITISH SCHOOL AT ROME.

still pavingstones in the fieldwalls. After following it E. for some 500 metres we reach a group of ancient buildings situated on the boundary line between the communes of Monte Porzio and Rocca Priora, marked by a fieldwall. First come the remains of a villa, with fragments of black and white mosaics, Roman Doric tufa columns, pieces of painted plaster, marble, etc. Late burials were also found here, and, I was told, a small bronze horse. On the platform is a reservoir 22 50 metres long and 2"5O metres wide ; and there are other chambers to the N. Below, on the W. side of the field wall, is a much ruined reservoir in two chambers, and close to it a deep circular shaft (marked W. in the map), which comes just at the point where the modern aqueduct which runs to Camaldoli terminates, and may be connected with it. To the S., a good way further down the hill, still just E. of the boundary line, is a reservoir with one chamber some 10 metres long, and below it, a little N. of the twenty-second kilometre stone of the modern road, the villa which no doubt it supplied, of which three apsidal rooms are preserved. In the ruins of the latter Professor Lanciani found the brickstamps C.I.L. xv. 565 k (about 123 A.D.) 1369 (164 A.D.—2 copies): the walling was partly in opus reticulatum, partly in brick, and partly in both together: in one wall the cubes of the former were placed horizontally : pieces of marble, a peperino column coated with stucco, and an Ionic capital in sperone were seen. To the E. of the uppermost villa is a large reser- voir in selce concrete measuring 2O'2O by io-8o metres outside, a plan of which is here given : the vaulting is quadripartite, and there are no external buttresses. It is referred to by Eschinardi (ed. Venuti, 1750), p. 272. The Staff Map marks ruins to the N. of the path at this point, which do not exist—either the path has changed its course since the map was made, or there is an error in the map. From here a branch path goes off, soon turning N., and descending sharply to S. Silvestro above Monte Compatri {supra, 382): there are no traces of antiquity along it, but it seems a necessary line of communication. The main path then bends slightly round the head of a valley, and here I noticed a rock-cut drain CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 407 o-47 metre wide and running S.E. Further S.E. are the remains of a villa in good brick work and opus incertum below the path; and above it (not marked in our map) is a villa with vaulted substructions in brick and opus reticulatum, on which may still be seen a plain white mosaic pavement with small tesserae, and fragments of fine marbles. Several red tufa columns about C40 metre in diameter, lie about there and look surprisingly fresh. From Atti del Camerlengato, Tit. iv. fasc. 246 we learn that on August 23rd, 1850, one Alessandro Mompieri, in searching for water for a new supply for the Commune of Monte Compatri, came upon some ruins on the south side of Monte Salomone (I imagine the reference must be to one of these villas, as there are no others answering to the description), with a fine coloured marble pavement, a fragment of a frieze of Parian marble decorated with ivy, various fragments of terracottas, of frescoes, stuccoes, and other decorations supposed to belong to a temple. Remains of a road were noticed leading up from the Via Latina to Monte Salomone, and other ruins were found on the upper slope facing S.W., one with reticulatum walls. The Commission composed of Jacobini, Grift, Visconti, and Canina, visited the site on September 19th, and stated the remains to be those of a villa. Besides the marble pavement, which was of giallo antico brecciato and bardiglio, they noticed a marble threshold on one side and beyond it the beginning of a black and white mosaic pavement. The remains of stucco, etc. showed that the building was of a certain magnificence: a terracotta antefix with a Scylla on it showed signs of the decline of art. The Commission then went to Tusculum and descended to Frascati, where they examined the statue of the Faun x in the shop of the chemist Signor Volpi. The Commune then obtained leave to continue the excavations. On May 5th, 1851, the president of the Municipal Commission of Monte Compatri wrote that so far he had not been able to spend more money on the excavations on Monte Salomone, which had been authorized by the Ministry, but that it had been reported to him that one Domenico Ciuffa had removed the squared blocks of stone which had been set aside, and had also devastated the opus reticulatum walls: the material he had placed in the lower story of his own house. Ciuffa, however, denied this, 1 I do not know which is the statue referred to. 408 THE BRITISH SCHOOL AT ROME. asserting that he had removed the stones from property of his own close to the hill in the course of clearing the ground. Beyond these two villas the path divides again, one branch descending gradually E.S.E. along the slopes. Its pavement is well preserved and can be traced as far as the point marked in the map, where it is interrupted by cultivation ; a few stones are preserved, but it soon dies out, and the course assigned to it in the map is entirely conjectural. The upper path is less certainly ancient, but must, I think, be taken as an almost necessary line of communication. Between the two (not marked in the map) I noticed a platform supported by walls of rough blocks of selce, and at the two places indicated in the map, S.W. and S. of point 725, are similar remains, with some brick debris at the latter; while a little further W., S.E. of the summit of Monte Salomone, is a vaulted chamber (substruction or reservoir) in selce concrete. Our path now reaches the road from Monte Compatri to Rocca Priora (the modern zigzag road follows a different course from the older path, which may represent an ancient line, though Stevenson (Vat. Lat. 10563, 23) was unable to find any trace of antiquity in the macchia) and also the short cut from the sixteenth mile of the Via Latina, which, again, perhaps, follows an ancient line. The village of Rocca Priora itself presents no traces of antiquity: there are no remains of ancient walls, either of a fortified enceinte or of later villas; and Tomassetti (294, n. 1) saw no other antiquities there but a fragment of a Christian relief. The mediaeval castle has been recently destroyed, and a new building constructed on the plan of the old one (ib. 295). Holste {Adnotationes ad Cluv. 780, 18) proposed to identify Rocca Priora with Corbio, an ancient city of Latium probably a member of the Latin League (Dionys. v. 61 Kopffivrwv). According to Liv. ii. 39. 4, Coriolanus after the capture of Lanuvium {Papers, iv. 4 n.) took Corbio, Vitellia, Trebium, Labici, and Pedum. That the list of these towns is given in any topographical order is by no means certain, and in any case Labici is the only one of which the modern representative (Monte Compatri, see Papers, i. 256 sqq.) is known to us. Of Vitellia we only know that it was a Roman colony in the territory of the Aequi (Liv. v. 29) and it does not appear in the list of members of the Latin League. Trebium, which is only mentioned in the present passage, can hardly be identified with CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 409

Treba, the modern Trevi nel Lazio, S.E. of Subiaco,1 and some authorities prefer to read Tolerium, which was one of the cities of the Latin League, and which is mentioned by Dionysius (viii. 17) as having been the first place captured by Coriolanus; Bola, Labici, Pedum, Corbio, Corioli, and Bovillae (c. 20) being reduced afterwards, and Lavinium unsuccessfully besieged.2 Thence he advanced to Rome; and then during a thirty days' truce marched against and took the other Latin cities (c. 36), Satricum, Ecetra, Setia, Polusca, Alba ? (in the text we find 'A\/3t»7Ta9) Mugilla, Corioli (?). Plutarch {Coriol. 28) [gives the order thus—Circeii, Tolerium,3 Labici, Pedum,4 Bola,5 Lavinium ; and does not name the other seven cities. Circeii is the first town mentioned by Livy; but both omit Longula, while Livy omits Bola, Bovillae, Ecetra, Setia, and Alba (?). We have further indications in the accounts of the wars against the Aequi, where we find (Liv. iv. 49) Bola and Labici in their possession towards the end of the fifth century B.C., and are told that the territory of these two places adjoined.6 Corbio, too, was obviously not far from Algidus (Liv. iii. 30)/ which plays so large a part in the wars between

1 There was, however, a farm called Trebia on the Via Latina in the Middle Ages (Tomassetti, 316). 2 I have suggested altering Lavinium to Lanuvium in Livy ii. 39 {inde L. recepit, a good deal earlier in the campaign : cf. Papers, iv. 4. n.) for topographical reasons ; and I think that the testimony of Dionysius and Plutarch is also in my favour: for both tell us that he raised the siege of Lavinium in order to march on Rome. 3 Tolerium has been identified with Valmontone by Nibby (Analisi, iii. 369) : see Papers, i. 273, n. 2. One of his arguments is the similarity of name with the Trerus or Tolerus or Tolenus (Sacco): cf. Bormann, Alllat. Ch'orogr. 78). In Analisi, iii. 22 he identifies Vitellia with Valmontone by a slip : for elsewhere (i. 466) he puts it at Civitella di Subiaco. 4 Of the site of Pedum we have no exact knowledge (Papers, i. 205, iii. 140), though Dessau (C.I.L. xiv. p. 288, n. 6) is not averse to placing it at Gallicano, as Nibby does (Analisi, ii. 551). The discovery in Tunisia of an inscription (Bull. Ant. de France, 1905, 177 : Bull. Com. 1905, 363) of a curator Viae Pedanae of equestrian rank does not help us in the matter of topography, for we do not know what this road may be. 5 I should think that Bovillae was the best reading here : for it is said to be not more than 100 stadia (12J miles) from Rome ; and this is very much too little, if, with Hiilsen (in Pauly-Wissowa, R.E. iii. 667) we are to place Bola in the upper valley of the Sacco : as he rightly remarks, a more precise identification is impossible. Nibby (Analisi, i. 291) placed it at Lugnano (now Labico)— see Papers, i. 273, n. 1. 8 Simul Aequos triennio ante accepta clades prohibuit Bolanis, suae gentis populo, praesidium ferre. Excursiones inde in confinem agrum Labicanum factae erant novisque colonis helium inlatum. 7 Horatius, cum iam Aequi Corbione interfecto praesidio Ortonam etiam cepissent, in Algido pugnat, multos mortalis occidit, fugat hostem non ex Algido modo sed a Corbione Orlonaque. Corbionem etiam diruit propter proditum praesidium. Cf. id. 27 and Dionys. x. 26. Of Ortona we have no further knowledge, except from Liv. ii. 43, and Dionys. viii. 91, from whom we hear of 410 THE BRITISH SCHOOL AT ROME.

Rome and the Aequi, and Ortona, of which we hear very little, must have been in the neighbourhood. Certainly Corbio was a strong position in those ancient days ; and Rocca Priora is the only site still occupied by a village which can claim to be as near Algidus as Corbio should be. We may therefore consider the identification a probable, but not a certain one; while as to Ortona, which seems to have utterly disappeared, we must confess our inability to fix it. In regard to Bola, Pedum, and Tolerium, we may content ourselves with saying that Nibby's identifications are probable, but not demonstrable in the present state of our knowledge, and that he is wrong in excluding Zagarolo from the list of possible sites for one of these ancient towns—he prefers to believe that it was only the site of an imperial villa. But, inasmuch as we have no real knowledge of the degree to which these towns or villages, which loom so large in the early history of Rome, were really fortified, and to what extent they would have left traces behind them, it is by no means necessary that they should have occupied sites upon which villages still stand. This is no doubt a more probable supposition (we so frequently find in Italy that the mediaeval town has returned to the primitive site on the hill, abandoning the Roman settlement in the plain : Labici itself, Privernum, Falerii, and many more are cases in point) and it also provides a far easier explanation for their utter disappearance. But it must be borne in mind that we are not bound to accept it. Thus in these cases we must, I think, be content with an approximate identification, until, as in the case of Labici, epigraphic or other evidence comes to our aid to settle the question. We should further note that the majority of the places mentioned in the account of the campaign of Coriolanus—Satricum, Bola, Corioli, Longula (?), Pedum, Tolerium, Vitellia—figure in Pliny's list of the 53 peoples of Latium which had utterly perished (N.H. iii. 68,69); and, though in some cases there is undoubtedly exaggeration in this expression, the fact is significant. The town of Carventum, one of the thirty cities whose inhabitants are mentioned by Dionysius (v. 61) among the members of the Latin League, and are perhaps identical with the Cusuetani mentioned by Pliny among a previous capture by the Aequi, and are told that it was a Latin city. Bunbury in Smith's Dictionary of Greek and Roman Geography, s. v. is inclined to accept the suggestion that we are to identify with the people of Ortona the Fortinei (?) of Dionys. v. 61 and the Hortenses of Plin. iii. 6g. This may or may not be so. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.-—III. 411 the peoples of Latium who had utterly perished, also appears in the warfare against the Aequi as an important post (Liv. iv. 53, 55, where the arx Carventana is mentioned). Nibby and others have failed to notice that it is expressly distinguished in the latter passage from Verrugo [infra, 424) ilia pro certo habenda, in quibus non dissentiunt [auctores) ab arce Carventana, cunt diu nequiquam oppugnata esset, recessum, Verruginem in Volscis eodeni exercitu receptarn—and have therefore identified it with Rocca Massima, whereas it should rather be placed in the neighbourhood of Algidus (cf. Hiilsen in Pauly-Wissowa, R.E. iii. 1628). From Rocca Priora a path descends first in an E.N.E. direction, and then due N., which falls into a road leaving the Via Labicana at the 18th mile, just W. of S. Cesareo {Papers, i. 266 Jin. and Map V.). The cutting where it turns N. at the ' cc' of 'Rocca' is fine, being about 10 metres deep and 3 wide, and may possibly be of ancient origin, though the rock is not hard ; but there are no further traces until we reach Fontana Chiusa, where there are a few possibly ancient pavingstones. I do not know what is the origin of the name Casa La Statua, and have not visited it. Just at the point where our path falls into that coming from S, Cesareo there is a piece of pavement in situ belonging to the latter : it is 2-40 metres wide and runs 200 E. of S. To the E. of the latter road I noticed a little brick debris. Another path which is probably ancient,1 as it is cut through the rock (though this in places is soft) and has what may be ancient paving- stones in it, leaves Rocca Priora in a S.E. direction and runs to the Via Latina. At the Fontanile Maggiore (from which is taken the view of Rocca Priora on Plate XXXVI. Fig. 1) it has a branch to the E. of which the antiquity is not. certain. To the E. of the summit of the Colle Trinciotto is a rough substruction wall of a single course of blocks of selce some 30 metres long, and another some 9 metres N.N.E. of it : on the slopes below is brick debris, and we probably have before us the remains of the terrace wall of some building. The path continues E., and is eventually joined by another branch which leaves the path from Rocca Priora lower down, passing N. of the Monte Fiore and of the Colle della Mola (the hill marked 640 metres above sea-level, N. of la

1 It is noted by Chaupy, Maison de Campagne d''Horace, ii. 164, who states (ib. 184) that it continued between Monte Compatri and Rocca Priora (supra, 382 n. 2), and that he noticed its cuttings and pavingstones. 412 THE BRITISH SCHOOL AT ROME.

Mola) where the path runs through a cutting some 2 metres deep and 2 wide. A little E. of the Colle della Mola, these two branches (supposing them to be of ancient origin) would have joined and fallen either into the road which descends N.E. along the Piano dei Mancinelli {Papers, i. Map VII.) or into that which runs due N. (it would fall just outside the E. edge of Map I. of the present volume—see the key-map given in vol. iv.) down the Valle della Giurnenta or Clementina {Papers, i. Map VI.), the prolongation of which is certainly ancient, as we have seen just above. I noticed ancient material used in a ruined house E. of the hill marked 522 in Map I. of this volume, and ancient debris and ruins both to the S. of and at point 331 {Papers, i. Map VI.), S. of the Casa della Pidocchiosa, and also to the E. of the last-named house, not far from the Via Labicana. Rosa in his map already alluded to indicates a number of ancient buildings in this neighbourhood, and these are no doubt the traces which he saw. The path down from Rocca Priora continues beyond these two branches in a southerly direction, crosses the Via Latina, and runs on until it reaches the Valle Vivaro {supra, 404). ' In a vineyard in the district of Tusculum ' said to lie towards the Valle della Molara, a vase was found with a hoard of gold coins of the late empire (474-565 A.D.). Lanciani saw twenty of them, all in mint condition —two of Zeno, eleven of Anastasius I., five of Iustinus I., two of Justinian {Not. Scavi, 1884, 192). The site is fixed as being the Colle Cappuccio or Colle Scapucciato, S. of Rocca Priora and W. of this path, by Tomassetti (p. 284, n. 2). Returning to the sixteenth mile of the Via Latina, we find, a little beyond it, on the S. of the road, the scanty remains of an outpost of the Castel di Molare, called I Muracci (Tomassetti, 297). The course of the Via Latina is traceable all the way along close to the modern path, as is indicated in the map: there is little paving in situ, but there are plenty of loose pavingstones to be seen. At the seventeenth mile it reaches its summit level (582 metres), which is no less than 221 metres (725 feet) higher than that of the Via Labicana at Mezza Selva {Papers, i. 270 ; iv. 8). The Casale Ostena Nuova is built actually upon the pavement of the road, as is so often the case. At the top of the descent E. of point 558, where it is crossed by the path from Rocca Priora, there are some remains of pave- CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 413

ment in situ in the modern lane; but in the straight line to the E. it is buried under the soil. Beyond the nineteenth mile the path through the Valle di Vivaro diverges to the S. {supra, 404); and just beyond it is the Lago della Doganella, considered by some authorities to be identical with Lake Regillus (see my papers cited supra, 321). I was wrong in denying that the lake was fed by springs, for these exist to the S. and supply the Fosso della Vetica (the water is slightly sulphurous). Indeed, on subsequent visits I have found it full of water (see Plate XXXVI. Fig. 2, for a view of it from the N.), but the other arguments against the identification are, I think, sufficient. To the E. of the lake the road ascends to the actual pass of Algidus, turning N. and then E. again : on the ascent is the twentieth mile: there are various "ruins here, but they are entirely mediaeval. They belong, I presume, to the Osteria dell' Aglio (Holste ad Cluv. 778, 25) or della Cava which was later on replaced by the Osteria Nuova. The road at point 560 reaches the summit of the pass by which it traverses the rim of the outer crater of the Alban Volcano between the Monte Castellaccio to the N. and the Monte Tagliente to the S. On both summits of the former (the western one appears in Map I., the eastern in Map II. of the present volume) there are extensive ruins, which are, so far as I can make out, entirely mediaeval, though there has been some use of ancient materials (I saw a granite column and some marble fragments): no doubt they were fortifications to guard the pass. They were seen by Chaupy {Maison de Campagne a"1 Horace, ii. 160), who believed them to be ruins of the temple of Fortune. Tomassetti, who himself saw and described these ruins (p. 300), is wrong, I think, in supposing that Chaupy alludes to the Colle Fiore ; while Gell {Topography of Rome and its Vicinity 43) is referring to the castle {infra, 414). On the Monte Tagliente, on the other hand, there are no remains at all, nor have I found anything ancient along the rim of the crater going southwards, until the Passo Brosciano, where N. of the path there is a concrete wall running W. by N. belonging to a small building^ apparently ancient. I have, indeed, ascended to every one of the many summits along the rim, in a vain search for the temples of Diana and Fortune (Horace, Od. i. 21. 1 Dianam tenerae dicite virgines . . . vos laetam fluviis et nemorum coma, quaecumque aut gelido prominet Algido . . . tollite laudibus; iii. 23. 9 quae nivali pascitur Algido . . . victima; iv. 4. 57 duris ut ilex 414 THE BRITISH SCHOOL AT ROME.

tonsa bipennibus nigrae feraci frondis in Algido. Carm. Saec. 69 quaeque Aventinum tenet Algidumque Diana. Liv. xxi. 62 supplicatio Fortunae in Algido (imperata). No temple of either deity is, it is true, actually mentioned, but one would think their existence must be assumed from these passages. The ruins I have seen, however, on the various hill-tops certainly cannot be attributed to either of these buildings. On the Maschio d'Ariano in particular, where the Staff Map (Map I. of the present volume) marks Algidum, there are, to my mind, no traces of anything earlier than the large mediaeval castle; nor can I agree with Chaupy (loc. cit.) and Rosa (map cit.) in seeing the traces of an ancient road ascending to it from the N. It seems to me clear that the remains spoken of by Gell (p. 43), who gives a sketch of what he calls the wall of Algidum, are simply the remains of this castle, the walls of which are largely built in small opus quadratum of brown granular tufa (the blocks being C30 to C44 metre high and 034 to 070 metre long) but contain mortar. Chaupy (Joe. cit.) and Nibby (Analisi, i. 121 sqq.) also visited the site, and Tomassetti gives a description of it and a planx (301 sqq.). I am sure he is right in suppos- ing that those authors who speak of remains of a temple of Diana here have mistaken for it the apse of the mediaeval church; and his supposition that this was the site of the temple of Diana, which would thus have been entirely obliterated by the mediaeval castle, has a good deal in its favour. The church was dedicated to S. Silvestro, and as an addition to Tomassetti's account of the mediaeval history of Algidus I may add that Duchesne {Lib. Pont. ii. 309, n. 57) refers to it the notice oppidum S. Silvestri in suam ditionem convertit {Paschalis ii.). Below the castle are various cuttings in the rock, probably forming the foundations of houses or huts, or perhaps simply quarries.2 The view from the castle is magnificent, and it is equally fine, though changing rapidly in details, as one follows the summits of the ridge to the S.W. over the Monte Peschio, the highest point of all (939 metres above sea- 1 Further excavation has laid bare more than he saw : thus, his conjecture as to the position of the gate has been verified, and the way in which it was defended made more clear : on the other hand, I could not find any subterranean chambers in the position which he assigns to them in the plan, but I saw two chambers just to the N. of the entrance running N. and S. There are also some rock cuttings (quarries ?) just to the E. of his ' roccia tagliata a picco.' 2 It may be a kindness to intending visitors to state that the Acqua Donzella has now been entirely appropriated to the supply of some village below the mountain, and that the wayfarer will hope in vain for any water there. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 415 level.1 I found no ancient remains until I reached the Maschio d'Arte- misio, 812 metres above sea-level, where there are scanty remains of mediaeval fortification to the E. and S.W. of the summit; while on the Monte Spina, 736 metres above sea-level, there is much brick debris. Immediately to the W. of this last summit is a deep depression in which runs the mediaeval post-road to Velletri {supra, 397), which crosses the rim of the outer crater at this point. Here, however, we are far beyond the limits of our maps, and we must now return to the Via Latina, which we left at the pass of Algidus. As we have already noticed {Papers, iv. 3 sqq.), the military importance of Algidus is especially prominent in the accounts of the warfare against the Aequi in 465-389 B.C., in which it was for a long while an advanced post of the latter, especially for their joint operations with the Volscians, though it formed no integral part of their territory (Hiilsen, s.v. Algidus in Pauly-Wissowa, R.E. i. 1476). There never was an actual town of Algidus or Algidum, as Dionysius thought x. 21 aicovcov irepl TTOXLV 'AXyiBbv adpoas elvat, ra? re OVOXOVCTKWV KCLL ra? AiKavoov 8vvd//,ei$ ; xi. 3 Al/cavol S' et? rrjv TovT€$ ofiopov oiaav ac^uat, Kal 7ro\Aa Brjaxxavref ainrj<; iv 'AA/yiSai nroXet, ridevrai rbv yapaica ; ib. 23, 28 he uses the phrase iv 'AXyiBai 7775 AlxavSu ^ftipa? : and Strabo (v. 3, 9, p. 237—97 Aart'vr].. . fiera^i) TOVO-KXOV Kal rov 'AXfiavov opovs Karnai eVi 'AXyiSbv troXi-^viov Kal TIiKTas is probably, as Hulsen thinks, alluding to a post station in the pass; while Stephanus of Byzantium (s.v.) is merely copying Dionysius. Many modern writers have followed this error, and have found a tower or castle of Algidum on Monte Fiore, where there are no ruins at all 2 (cf. Nissen, Ital. Landeskunde, ii. 595 fin.). Livy invariably uses the phrase in Algido, which Ovid (Fasti, vi. 721) paraphrases as Algida terra. In imperial times we hear of this locality as a favourite resort; cf. Statius, Silv. iv. 4. 16 hos Praeneste sacrum, nernus hos glaciale Dianae, Algidus aut horrens aut Tuscula protegit umbra, Tiburis hi lucos Anienaque frigora captant: Martial x. 30. 6 mentions Tusculanos Algidosve secessus among places usually selected for villas : Silius Italicus (xii. 536) calls the place

1 The Fontanile Moscaccio, some way down the S.S.E. slope of this mountain, is fed by a good spring, brought by a rock-cut channel o'6o metre wide, possibly of Roman origin. 2 Tomassetti knows nothing of the destruction of the castle by Sixtus V., which is mentioned as a tradition by Holste (loc. cit.). 416 THE BRITISH SCHOOL AT ROME. amoena Algida and Symmachus (Ep. iii. 50) refers to it as a summer resort—ut aestas mihi Praenestino Algido frangeretur, while Pliny speaks of chick peas (ciceres) and radishes (raphani) as growing especially well in Algidus (H.N. xviii. 130, xix. 81). Some later writers (e.g. Hieron in Chron. Euseb. a. Abr. 1528, Eutrop. i. 17, 18, Auctor de vir. ill. 17) speak of Algidus mons. (There is, I think, no doubt that the 'A\ya>8a>v of Procopius (B.G. iii. 22) refers to Alsium.) We have seen that the remains of villas in the neighbourhood of the pass are somewhat scanty, and the name must in imperial times have taken a somewhat wider sense, referring especially to the northern portion of the rim of the outer crater : for in the southern part there are, so far as I know (though, owing to the existence of forests, exploration is not easy), far fewer ancient buildings.

XXI.—THE VIA LATINA FROM THE TWENTIETH TO THE THIRTIETH MILE. Immediately after traversing the pass, the Via Latina begins to descend ; biit, before it does so, a branch goes off to the left, which reaches the Via Labicana on the Colle Tre Are (Papers, i. 274: cf. Map. VII.). The pavement is preserved along the left-hand track, until it takes a sharp bend S.: here, in a field, where V is marked in the map, is brick debris, and remains of fine coloured marbles. At the Casale di Mezza Selva itself there is nothing ancient: beyond it the road is crossed by that mentioned in Papers, i. 270, which Holste (Cod. Dresd. F. 193, 72V) notes as ancient— unless he is referring to that near II Fico1 (cf. infra, 417)—and beyond again are further traces of paving. According to Foggini (Mus. Cap. iv. p. 24) the cippus bearing the inscription C.I.L. xiv. 3030 (cf. Matz-Duhn 3879) was found under his own eyes in the district called Mezza Selva, three miles (sic) from Palestrina ; while according to Marini (Vat. Lat. 9131, I2V) it was discovered at the Osteria della Cava (supra, 413) in 1768, whereas Foggini (ap. Amaduzzi, Anecd. Hit. iii. (1774), p. 468, n. 13) states that it was found in 1770. At point 403 (the district is known as II Fico) Stevenson marks a road going due S. in his map, probably just along the boundary line, and ascending to the Colle della Regina and Colle della Castagna (infra, 417): 1 What he means by referring to a paved road which crosses that which we are following half a mile before the Casale, coming from the Labicana, and going to the Piano di Velletri, I do not understand. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 417 there are no pavingstones now visible on this portion of it, but to the N. of the main road there is pavement, which is soon lost in a thicket of broom. We may probably suppose therefore that it ran on to the Via Labicana and reached it at the Valle dei Gelsi. A little to the E., on the N. of the road we have been following, is a reservoir and some debris, indicated by Stevenson ; and further on again on the S., before the descent to the Via Labicana begins, is the brick debris of another villa. For the Via Labicana cf. Papers, i. 274. Returning to the pass, and following the Via Latina itself (Via di Artena in the map) we find ancient pavingstones in it, in many cases relaid, and then the stones may be seen in a ploughed field to the N. On the N. edge is a concrete reservoir with a single chamber ; and further on, just before reaching the Casa Borghese, are remains of a building in opus quadratum of tufa, opus reticulaturn of selce, and brick, and of a long narrow reservoir. At the Casa Borghese are two Corinthian capitals, a female draped statue, of white marble, lacking the head, which was in a separate piece, also part of what seemed to be the torso of an animal. The modern path then passes through a cutting which is not ancient: at the twenty-second mile, on the S. side of, and above it, are four masses of fallen concrete, the remains of a tomb : the course of the road is traceable for the whole way, but only at intervals is there pavement in situ It continues to descend gradually, until it reaches, at the point called Le Crocette, a path which crosses it at right angles. On the descent to the N. there are a few loose pavingstones, and in the valley the path divides, one branch going towards Casale di Mezza Selva and forming a part of the line of road described in Papers, i. 270. There is little evidence for the antiquity of this portion, but we are perhaps justified in assuming it as a necessary link in the chain, and the map opp. p. 90 inserted in Fabretti's De Aquis, ed. 1788,1 marks it as ancient (via in Latinam et Appiam tendens). The other branch going N.E. can be traced by pavement in situ as marked in the map: it first ascends E.N.E. along the slopes of the Colle della Castagna, then runs due N. up a gully, at the head of which are scanty foundation walls, and to the E. of them a reservoir ; while, further E.N.E., on the Colle della Regina above the head of a lateral valley going E., are the remains of an extensive building. Loose pavingstones may be

1 It is clear from a passage in his Inscriptiones (p. 415) that this is the map originally inserted in his Apohgema Iasilhei in Grunnovium, a rare tract directed against Gronovius which I have never seen. E E 418 THE BRITISH SCHOOL AT ROME. seen all about; and on the descent towards the Spallette di Piovito and the Valle Carbone the pavement is in situ, running N.E. for a while and then almost due N. At the bottom of the descent it would fall into the path running N. from the Fontanile delle Macere along the bottom of a broad valley between the Colle della Castagna and the Colle del Rapiglio, which is probably the road alluded to by Holste {ad Cluv. 949, 33) as coming from the Via Latina to the Via Labicana at the Colle dei Quadri and by Chaupy {op. cit. iii. 463 ; cf. Papers, i. 274), though there are now no traces of it to be seen. To the S. of Le Crocette there are loose pavingstones and some pavement in situ along the path as far as the edge of the wood ; but along the Valle Ontanese there are no traces of it to be seen as far as the Via Ariana, and the slight depressions in which the track runs in places cannot claim to be of ancient origin. Further S. too it presents no traces of antiquity, though if followed due S. it eventually reaches the Via Doganale near the Lake of Giulianello ; and Holste {Cod. Dresd, F. 193. 74) says that he was told that it was paved in his day.1 That the portion of the Via Ariana to the S.W. of this point is of ancient origin is clear from the testi- mony of De la Blanchere—Chapitre de I'histoire pontine 48—from Sav. e'tr. a VAcad. des Inscr. et B.-L. x. i. (1889)—who states that there were traces'of it near Fontana di Papa, S.W. of l'Ariano, and by the discovery of the pavement in situ on the line of the present road further S.W., about three kilometres before reaching Velletri, in a locality called ' Pietreliscie,' a name which recalls the time when the modern road was not yet made, and the ancient road was still in use {Not. Scavi, 1899, 338). In 1834, the contractor for the road, in digging for material in land belonging to the Municipality of Velletri, at a place called II Macchione, within the territory of Ariano, came upon some walls of inferior con- struction, probably mediaeval, and some fragments of marble columns, etc. The report that a statue was also discovered proved to be untrue {Atti del Canierlengato, Tit. iv. fasc. 2223). To the E. of Le Crocette, both N. and S. of the road, there is brick debris. To the E. of this Stevenson marks an ancient road, running right along the top of the hill down to the Via Ariana : there are no traces of it, 1 Disse un tale mastro Antonio Carlone di Montefortino che ad un certo luogo detto l'Uscione si rincontra un' altra selciata che tira a man dritta verso S. Juliano . . . . e tengo io per sicuro .... che sia quella medesima che traversando la Latina passa vicino all' hosteria di Mezza Selva e si rincontra con la Labicana a un certo luogo detto La Torre {supra, 418). CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.^—III. 419 and I cannot help thinking that he has marked wrongly the road of which we have just spoken. On the other hand he is right in indicating that the ancient road kept to the N. of the modern track, from the twenty-third mile onward, and did not follow the narrow defile by which this descends (in which there is mediaeval paving) but kept straight on, over the end of the hill, some 100 metres further N., preserving thus a straighter line than I had given it. We now reach the Fontanile delle Macere in the valley, which is bordered all round with pavingstones. Cultivation has destroyed unfortunately all trace of the ruins of which Chaupy speaks (ii. 171; iii. 463) as being so plentiful. He remarks rightly that the situation was not a favourable one for a post station, but I think that Ad Pictas must be placed here {Papers, i. 275) or close by. The steep curving ascent, too, beyond the fountain, is probably not ancient: after this, though the line of the road is for a while lost in the vineyards, he marks it a little further S. than I have done—perhaps not correctly, for there are two parallel walls, possibly belonging to a tomb near the top of the ascent, and then some loose pavingstones, which are probably not far from their original position. Just beyond the twenty-fourth mile the Via Latina is cut at right angles by a road coming from the Via Labicana at the Colle della Strada,1 which has no pavement in situ but is in places deeply cut {Papers, i. 274) and which at this point falls into the Via Ariana {supra, 418) ; and I feel sure that this must be the road to which Chaupy refers (ii. 171 ; iii. 463) as cutting the Via Latina at right angles immediately after the Fontanile delle Macere—especially as the portion of the Via Ariana to the E. of this point is not ancient, and indeed, when the Via Latina was in existence, would have had no raison d'etre. On the S.W. slope of the Colle Rotondo {Papers i., Map vii.) is a rect- angular open piscina, and to the N. traces of a villa, and further N. again of another. The Colle Casalupi is by some authors, and very likely correctly, thought to be identical with the Fundus Casaluci of the inscription of SS. Giovanni e Paolo (Fabretti, Inscr. p. 416), which with the Fundus Stagnis is placed at the thirtieth mile of the Via Latina. Marocco, Stato Pontificio ix. 39 notes ruins there. The road passing E. of the Colle Rotondo is probably also of ancient origin {Papers, i. 274), and though further

1 It goes onN. to Labico, though Westphal (p. 76,/K. ), whom Kiepert must be following (map to C.I.L. xiv), prolongs it to Valmontone. E E 2 420 THE BRITISH SCHOOL AT ROME. investigation has shown me that there are, as a fact, no traces of ancient pavement on the Via del Buon Viaggio and the Via Doganale, I still believe this line to be of ancient origin. This is the Via Corani tendens of Fabretti's map. The tomb on the S. of the Via Latina just to the E. of the Via Ariana is that which was discovered in 1890 {Papers, cit.). On the further side of the railway, to the N. of the road, is a reservoir in limestone concrete (for we are now leaving the volcanic region of the Alban Hills, and reaching the limestone district of the Volscian mountains), while S. of the Colle Pastore and on the summit of the Colle Monte Crepi are scanty remains of probably ancient buildings. I have been able to take a rubbing of the milestone, of which I published a copy in Papers, i. 278, n. 2, from which it seems to me clear that the reading should be as follows: IMF INVICTO MAXEN AV PERPETV XXV We have thus the twenty-fifth or twenty-sixth milestone of the time of Maxentius. Whether it had really been found at La Pescara I do not know ; but in any case there is no doubt that it belonged to the Via Latina. Beyond the twenty-fifth mile, on the N. side of the road, is a small rectangular building—a tomb or a reservoir. Traces of pavement on this portion of the road are scanty, and to the E. of the divergence of a path to the E.N.E. past S. Egidio, which reaches S. Ilario by way of the Valle Materna, and probably, but not certainly, follows an ancient line (cf. Papers, i. 279), there are no traces of antiquity at all along the Via Latina, which has completely disappeared in the vineyards. Marocco, op. cit. 38, saw fragments of columns in the Valle Rapello (probably the same as the Valle Stefano) below the Colle di S. Egidio, and the ruins of a villa with mosaic pavements and hypocausts on the Colle Cadellino. To the N., W. of the so-called Grotta di Serracina, are the remains of a villa: the Grotta itself we could not find, nor does the path along the N. side of the Colle dci Lepri appear to be ancient. E. of the Casa Guglielmetti, just at the approach to the modern village of Artena, there is a tomb of white lime- stone concrete, which may mark the line of the road, but still no paving- CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 421 stones are visible, nor are there any in the straight section E. of Artena, though there is a little mediaeval paving. The scanty antiquities of Artena itself, and the interesting remains of La Civita (an ancient town to which we cannot give a name) are fully dealt with by Dr. G. J. Pfeiffer and myself in a paper in the Sitppleinentary Papers of the American School, i. (1905) 87 sqq. There is another milestone {C.I.L. x. 6884) belonging to this part of the road and now preserved at Artena, the number on which is unfor- tunately not clear. Serangeli, in his MS. account of Montefortino (see C.I.L. x. p. 591) gives the following account of its discovery: 'The Via Latina crosses the vineyards in the district of Le Valli opposite to Monte- fortino .... near the property of the rural benefice of S. Pietro, cutting the modern road to Valmontone, where was found lying on the ground the milestone recognised as such by Fabretti and Adrien Auronet (more correctly Auzout, as Fabretti gives it), a Frenchman. It is now in the courtyard of the Palazzo Borghese (where it still is) having been trans- ported there in 1698 by my orders, I having been urged to do so by the above-mentioned antiquaries.' Fabretti {Inscriptiones, 414, 367) gives a copy of the inscription on the milestone, which he believes to have borne the number 34, and states that it was found two miles before the station of Ad Bivium, which he rightly puts at S. Ilario. He accepts, however, the distance of thirty-five miles from Rome by the Via Latina given by the Antonine Itinerary, which is excessive. Stevenson read the text thus Diomino) N(ostro) Imp{eratori) Caes{ari) Cl{audio') Iulia\no~\ P\io felici Auguslo X~\ XXIIII. I could not make out the number at all when I saw it; but the spot at which it was found must have been very near the twenty-seventh mile, as will be seen from our map. Fabretti's reading may well have been influenced by his mistaken idea as to the length of the road : Stevenson confesses to having had some doubt about the number, and I therefore think it legitimate to suggest that this may be the twenty- seventh milestone : the number would thus be complete, and we only have to suppose that XXVII has been misread as XXIIII, which, with an inscription of the time of Julian the Apostate, is not difficult. Marocco {op. cit. 40) notes the discovery here of a cinerary urn of palombino, with a gold necklace decorated with precious stones, and of the sepulchral inscription C.I.L. x. 5986. In this, however, he may not be correct, for the latter had been found over a century before. 422 THE BRITISH SCHOOL AT ROME.

After the twenty-eighth mile the road divides, one branch going on eastwards along the foot of the mountains, while the other turns N.E. and runs, first ascending and then descending again, to join the Via Labicana at S. Ilario: on the N. side of the latter, near the twenty-ninth mile, is a large reservoir, and further on, on the Colle Maiorana, a villa ; and along this branch are many loose pavingstones. I have not thought it worth while to carry the map further E., as the junction of the two roads is shown in Papers, i. Map VIII., and I have nothing to add to it. I still feel no doubt that S. Ilario is to be identified with Ad Bivium. As I have pointed out {Papers, i. 280; iv. 7), the distances by the two roads are so nearly identical that Mommsen's arguments for the priority of the Via Labicana would fall to the ground ; and the military import- ance of the pass of Algidus leads me to suppose the existence of a road to it in very early times, though I certainly think that at a later period the Via Labicana became the more important road.1 It may be interesting to quote Holste's account of the catacombs of S. Ilario {Cod. Dresd. f. 193, 73V): ' S. Ilario is a hill, on the top of which are seen traces of an oratory, and below a very fine ancient cemetery, all cut out in the hard tufa. The entrance faces N.W., and at the entrance on each side there opens a passage : that on the left ends in a niche, that on the right turns and leads into other passages which go out on the same side. In the middle, on the right, one enters by a low opening into a small square room, where there is a little water: they call it the fountain of S. Ilario, the water from which the women drink from dead men's skulls in order to have milk ; and of this water they relate many wonderful things, and also of the skulls.' The castle of Piombinara or Pimpinara {Papers, i. 280), on the op- posite side of the Sacco from the railway station of Segni, is now fully described, with a plan, by Tomassetti (Vie Labicana e Prenestina, Rome, 1907, 118). The lofty tower, one of the highest in the Campagna, com- manded a view over both the Via Labicana and the Via Latina, and the castle occupied a hill in a narrow part of the Sacco valley, where there is now only just room for the road and railway to pass, one on each side of the stream, close to the point where the Via Labicana must have crossed the

1 I should add that I have, since I wrote Papers, iv. 8, n. followed the high-road from Teano to Cassino : I could not see any certain traces of antiquity along it, but this does not prove that it is not of ancient origin, and the line seems to me the natural one. I do not feel sure about the object or antiquity of the building I had thought to be a tomb. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 423

Sacco, and been joined by the road along its valley, vjhtch is now followed by the modern road from Valmontone. A view of the castle from the S.E. is given on Plate XXXIV. Fig. 2. The original line of the Via Latina is, if our view is correct, that which kept on due E. from the bifurcation between the twenty-eighth and twenty-ninth mile, and joined the line of the Via Labicana at Compitum Anagninum, or perhaps originally even further on ; but we shall find it to be tortuous and difficult compared with the Via Labicana, so that it must naturally have become of minor importance. Just before the twenty-ninth mile the road crosses a stream descending from the S., the upper valley of which is known as the Vallone Grotta Ferrata. Some way up it is a hill called the Colle l'Oppi, and this is probably the site of some discoveries recorded in Atti del Camerlengato, Tit. iv. fasc. 2209. In August, 1834, a shepherd named Giovanni Villani discovered a tomb formed of tiles, nineteen in number, each two palms by one (C446XC223 metre) without stamps: within it was the skeleton of a child, with a gold necklace decorated with dark oval stones fixed to it, three gold rings, a ' ciappetta' of gold with a violet stone in it and two gold hooks at the back, and some pieces of glass. Further search led to no discoveries, but the ground was covered with debris. The objects were bought for the Papal museums for 16 scudi 20 baiocchi. The locality is defined as being in the Quarto degli Oppi, in the district called La Suattera, four miles from Montefortino, five from Colle Ferro, and seven from Piombinara. Just after this, on the extreme S. edge of Map VIII. of Papers, L, is a small reservoir of limestone concrete, while to the S. on the hill are the so-called Muracci di Crepadosso, the remains of another reservoir. We then cross the boundary line, and enter a small isolated portion of the territory of the Commune of Rome. The rest of the course of the road, as far as Compitum Anagninum, is shown in the Staff Map of Italy, fo. 151, iii. (Anagni), 1 : 50,000. The road now turns S.E.: on its S.W. side is the brick debris of a villa, just under the 'Cr' of 'Crepadosso,' while to the S. of Fontana Viola, where are the letters 'di,' are the remains of three reservoirs in limestone concrete, with much debris lying about. To the N.E. are remains of a villa, N.W. of the Casale Colle Ferro, N. and E. of which are other reservoirs. Nibby {Analisi, iii. 473) and others have placed Verrugo 424 THE BRITISH SCHOOL AT ROME. here: it appears to. have been a locality (whether a town or merely a strong military position is uncertain) in Volscian territory, the possession of which was frequently disputed in the warfare between the Volscians and the Romans in 445-394 B.C. ; but we cannot certainly localize it, and the site of Colle Ferro has perhaps hardly sufficient military import- ance. (In Liv. iv. 55 it is spoken of as in the Volscian territory, whereas in iv. 1 we find laeti audiere patres . . . Vulscos Aeqnosque ob conimunitam Verruginem fremere, and in v. 28 it is spoken of in connection with the war with the Aequi: so that it must have been on the frontier between the territory of both peoples and that of Rome. Similarly, in one passage Diodorus (xiv. 11) speaks of it as a Volscian city, in another (ib. 98) in connection with the war against the Aequi.) It is therefore unlikely that it is to be identified with Rocca Massima; and we have seen above that the traditional view, that the arx Carventana is to be placed there, will not hold good {supra, 411). There is a large reservoir to the S.E. of Colle Ferro at point 222, and a small one in the fork of the roads to Segnix and to Gavignano, close to which is a tomb: to the N. of these, close to the latter road, is a villa with substruction walls in polygonal work and opus incertum ; and there is another villa a little to the E. at point 232, with more ruins again S. of it. Two groups of ruins marked in the map to the N. of the Via Latina are both reservoirs. The frequency of these is remarkable, and it is clear that the water supply of the district cannot have been good. The Via Latina in the meantime has no traces of antiquity (except that the mediaeval ruins W. of the Ponte del Pisso contain a few paving- stones), and the pavement and all the bridges are mediaeval. The path going N. a little before I Rossili passes through a cutting which appears to be of ancient origin, and would form a line of communication between the Via Labicana and Segni. Beyond this the path undulates considerably and follows a tortuous course. By the chapel marked ' Cona fra Rocco' in the map there is an ancient reservoir or tomb in concrete, but otherwise there are no traces of antiquity. We cross the Sacco close to the railway station of Anagni: from that point, though I have not followed it, I suppose that the Via Latina would have run N.E. to the Osteria della

1 The bridge of the former road over the Sacco (now replaced by a modern iron one) was ancient in Fabretti's day {Inscriptiones, p. 416—cf. Marocco op. cit. ix. 43) but has no doubt been carried away by the frequent floods to which this river is subject. CLASSICAL TOPOGRAPHY OF THE ROMAN CAMPAGNA.—III. 425

Fontana (Compitum Anagninum) {Papers, i. 281). The distance would have been about forty-one miles from Rome, which would only agree fairly well with the itineraries which give the distance from Rome by the Via Labicana as forty miles, and from Ad Pictas as fifteen miles, which is too little (id. 216). Due S. from the Osteria della Fontana an ancient road must have run to Villamagna on the S. side of the Sacco, and on the N. slope of the Volscian hills. This we know from the duplicate inscriptions C.I.L. x 5909 (cf. the Addenda, p. 982) still preserved at Anagni, set up by Sept- imius Severus and Caracalla (Geta's name has been erased) in 207 A.D., recording that they paved the via quae ducit in villam magnani. The name La Selciatella, which may be seen on the map a little N. of the railway, is significant, though I do not imagine that any pavement is still preserved along its course, inasmuch as Holste, who, like Marangoni, saw it in a good state of preservation, tells us that its destruction was being begun in order to provide material for the new one. I have not yet visited Villamagna ; but the remains of a large villa there are described by Ambrosio de Magistris (Storia di Anagni, Rome, 1889, i. 198). That it belonged to M. Aurelius, however, is by no means certain. From the Atti del Camerlengato, Tit. iv. fasc. 599, we learn that on April 2, 1827, the Cardinal Chamberlain wrote to the Delegato of Frosinone, informing him that he had heard that excavations had been carried on without leave at Villamagna. The result of inquiries made was that a letter of recommendation to the Bishop of Ferentino, given by the Cardinal Chamberlain himself to the sculptor Laboureur, who desired to make these excavations, had been taken by the Chapter of Anagni to amount to a permission, and that the excavations, conducted in September, 1826, had led to no result.